The Mid Market CEO Playbook
The Mid Market CEO Playbook
The Mid Market CEO Playbook
Innovation Playbook
Actionable strategies for mid-market leaders to unlock AI-powered growth through focus, alignment, and disciplined execution.
1. Introduction
A systematic approach to identifying, implementing, and scaling AI initiatives that transform how your organisation creates, delivers, and captures value
This playbook isn't about technology—it's about business transformation. While others focus on AI features and capabilities, we start with what matters most: your business model and how AI can fundamentally change it for the better.
Built on the proven Business Model Canvas framework from Strategyzer, this guide helps you apply their nine building blocks specifically to AI transformation. Whether you're running a startup exploring your first AI implementation or leading an established enterprise looking to systematise your approach, this playbook provides the frameworks and methodologies you need to succeed.
The difference? We don't begin with "What can AI do?" We start with "How does your business work today, then what can AI do, and then workshop where could AI create the most value?" This business-first approach ensures every AI investment connects to real customer outcomes and sustainable competitive advantage.
Download the Strategyzer Business Model Canvas here.
1. Introduction
A systematic approach to identifying, implementing, and scaling AI initiatives that transform how your organisation creates, delivers, and captures value
This playbook isn't about technology—it's about business transformation. While others focus on AI features and capabilities, we start with what matters most: your business model and how AI can fundamentally change it for the better.
Built on the proven Business Model Canvas framework from Strategyzer, this guide helps you apply their nine building blocks specifically to AI transformation. Whether you're running a startup exploring your first AI implementation or leading an established enterprise looking to systematise your approach, this playbook provides the frameworks and methodologies you need to succeed.
The difference? We don't begin with "What can AI do?" We start with "How does your business work today, then what can AI do, and then workshop where could AI create the most value?" This business-first approach ensures every AI investment connects to real customer outcomes and sustainable competitive advantage.
Download the Strategyzer Business Model Canvas here.
1. Introduction
A systematic approach to identifying, implementing, and scaling AI initiatives that transform how your organisation creates, delivers, and captures value
This playbook isn't about technology—it's about business transformation. While others focus on AI features and capabilities, we start with what matters most: your business model and how AI can fundamentally change it for the better.
Built on the proven Business Model Canvas framework from Strategyzer, this guide helps you apply their nine building blocks specifically to AI transformation. Whether you're running a startup exploring your first AI implementation or leading an established enterprise looking to systematise your approach, this playbook provides the frameworks and methodologies you need to succeed.
The difference? We don't begin with "What can AI do?" We start with "How does your business work today, then what can AI do, and then workshop where could AI create the most value?" This business-first approach ensures every AI investment connects to real customer outcomes and sustainable competitive advantage.
Download the Strategyzer Business Model Canvas here.
2. Making AI Happen
Most organisations struggle with AI because they lack a clear starting point. Most organisations I have worked with have ‘shiny new thing’ syndrome; their catalyst is fear of missing out, or they have seen or heard of something and they have that buyers itch. This focused sprint changes that, taking you from confusion to commitment with a shared understanding of your business model and your most promising AI opportunities.
Why This Approach Works
We start with your business model, which is something many organisations don’t fully understand anyway, especially as they grow and scale. We help you understand how you create, deliver, and capture value today, and identify where AI might fundamentally change those dynamics for the better.
This outcomes-first approach delivers three clear advantages:
Every AI initiative connects to measurable business outcomes rather than becoming an expensive science project that impresses technical teams but delivers no real value.
You identify transformative opportunities others miss because you're not just thinking about AI as automation, you're thinking about it as business model innovation - you are thinking of it as a way to completely transform how you deliver value to your customers.
As a result, technical and business teams speak the same language, eliminating the communication breakdowns that kill most AI projects before they deliver results.
What You'll Achieve
By the end of this sprint, you’ll achieve what most take months to do:
Shared Business Model Clarity: Everyone involved understands exactly how your organisation operates today using a proven framework. This becomes your foundation for identifying where AI creates the most value.
3-5 Ranked AI Opportunities: Not just ideas, but proper initiative briefs with clear business justification, technical feasibility assessment, and risk evaluation. Each one ready for investment decisions.
One Committed Pilot — A selected initiative with clear metrics, data sources, governance, and cost limits. You’ll know when to scale or stop — saving time, money, and reputation.
The Team You Need
AI success needs cross-functional collaboration, not large teams. The most effective approach blends internal expertise with targeted external support — often with overlapping roles:
Product Owner defines success and ensures customer value.
Tech Lead grounds ideas in technical feasibility and data reality.
Risk Lead manages compliance, bias testing, and governance.
Data Owner understands data quality, access, and limitations.
Finance Partner keeps focus on unit economics and cost control.
Delivery Manager drives momentum and accountability across teams.
2. Making AI Happen
Most organisations struggle with AI because they lack a clear starting point. Most organisations I have worked with have ‘shiny new thing’ syndrome; their catalyst is fear of missing out, or they have seen or heard of something and they have that buyers itch. This focused sprint changes that, taking you from confusion to commitment with a shared understanding of your business model and your most promising AI opportunities.
Why This Approach Works
We start with your business model, which is something many organisations don’t fully understand anyway, especially as they grow and scale. We help you understand how you create, deliver, and capture value today, and identify where AI might fundamentally change those dynamics for the better.
This outcomes-first approach delivers three clear advantages:
Every AI initiative connects to measurable business outcomes rather than becoming an expensive science project that impresses technical teams but delivers no real value.
You identify transformative opportunities others miss because you're not just thinking about AI as automation, you're thinking about it as business model innovation - you are thinking of it as a way to completely transform how you deliver value to your customers.
As a result, technical and business teams speak the same language, eliminating the communication breakdowns that kill most AI projects before they deliver results.
What You'll Achieve
By the end of this sprint, you’ll achieve what most take months to do:
Shared Business Model Clarity: Everyone involved understands exactly how your organisation operates today using a proven framework. This becomes your foundation for identifying where AI creates the most value.
3-5 Ranked AI Opportunities: Not just ideas, but proper initiative briefs with clear business justification, technical feasibility assessment, and risk evaluation. Each one ready for investment decisions.
One Committed Pilot — A selected initiative with clear metrics, data sources, governance, and cost limits. You’ll know when to scale or stop — saving time, money, and reputation.
The Team You Need
AI success needs cross-functional collaboration, not large teams. The most effective approach blends internal expertise with targeted external support — often with overlapping roles:
Product Owner defines success and ensures customer value.
Tech Lead grounds ideas in technical feasibility and data reality.
Risk Lead manages compliance, bias testing, and governance.
Data Owner understands data quality, access, and limitations.
Finance Partner keeps focus on unit economics and cost control.
Delivery Manager drives momentum and accountability across teams.
2. Making AI Happen
Most organisations struggle with AI because they lack a clear starting point. Most organisations I have worked with have ‘shiny new thing’ syndrome; their catalyst is fear of missing out, or they have seen or heard of something and they have that buyers itch. This focused sprint changes that, taking you from confusion to commitment with a shared understanding of your business model and your most promising AI opportunities.
Why This Approach Works
We start with your business model, which is something many organisations don’t fully understand anyway, especially as they grow and scale. We help you understand how you create, deliver, and capture value today, and identify where AI might fundamentally change those dynamics for the better.
This outcomes-first approach delivers three clear advantages:
Every AI initiative connects to measurable business outcomes rather than becoming an expensive science project that impresses technical teams but delivers no real value.
You identify transformative opportunities others miss because you're not just thinking about AI as automation, you're thinking about it as business model innovation - you are thinking of it as a way to completely transform how you deliver value to your customers.
As a result, technical and business teams speak the same language, eliminating the communication breakdowns that kill most AI projects before they deliver results.
What You'll Achieve
By the end of this sprint, you’ll achieve what most take months to do:
Shared Business Model Clarity: Everyone involved understands exactly how your organisation operates today using a proven framework. This becomes your foundation for identifying where AI creates the most value.
3-5 Ranked AI Opportunities: Not just ideas, but proper initiative briefs with clear business justification, technical feasibility assessment, and risk evaluation. Each one ready for investment decisions.
One Committed Pilot — A selected initiative with clear metrics, data sources, governance, and cost limits. You’ll know when to scale or stop — saving time, money, and reputation.
The Team You Need
AI success needs cross-functional collaboration, not large teams. The most effective approach blends internal expertise with targeted external support — often with overlapping roles:
Product Owner defines success and ensures customer value.
Tech Lead grounds ideas in technical feasibility and data reality.
Risk Lead manages compliance, bias testing, and governance.
Data Owner understands data quality, access, and limitations.
Finance Partner keeps focus on unit economics and cost control.
Delivery Manager drives momentum and accountability across teams.
3. Your Step-by-Step Guide
Step 1: Foundation and Goal Setting
Begin by defining exactly what your business aims to achieve — growth, margin improvement, risk reduction, or a better customer experience. Be specific with targets and timeframes. This isn’t about AI yet; it’s about understanding what your business delivers and what success looks like. Focus on outcome-based metrics that matter, such as “Reduce customer service cycle time by 30%” rather than “achieve 95% model accuracy.” Set thresholds for what counts as real improvement.
Next, introduce the Business Model Canvas to map how your organisation creates, delivers, and captures value. You and your team will use it to identify where AI can make a meaningful difference and to assess the impact of any changes. This shared structure becomes the foundation for every AI conversation that follows.
Step 2: Current State Reality Check
Now, make sure your business model canvas is complete. This means gathering real data on how your business operates today. Have someone walk the floors, document key processes, customer journeys, and decision points — including timing, handoffs, and failure modes. Add metrics wherever possible. Understanding how customers actually interact with you reveals where automation and business model improvements may lie.
Next, create an inventory of your systems, data sources, and integrations. Be brutally honest about data quality, access restrictions, and technical debt — most AI projects fail because these issues surface too late. Finally, document your cost structure, revenue streams, and pain points — including customer friction, inefficiencies, and regulatory or technical constraints. This becomes your foundation for evaluating where AI can drive real, sustainable value.
Step 3: AI Opportunity Discovery
Work through each of the nine canvas areas systematically, asking how AI might change the dynamics in each area. Whilst its tempting to leave this to one person to do, many minds are better than one. This works best when a diverse range of people from across your organisation get together in a room. The person who did the legwork in understanding your business in Step 2 should present, the factual accuracy should be agreed, and then start to brainstorm ideas.
That can be the hard bit, and the format of the workshop needs to be carefully managed to help this. Understanding who you have in the room, and how to get the most out of them is critical. There are plenty of good techniques that you can use to effectively brainstorm.
Consider inviting key customers or suppliers to join for an outside perspective.
For each opportunity, create a one-page brief focused on the business problem first, then how AI solves it – clearly defining what it will and won’t do. Aim for three to five substantial opportunities rather than a long list of minor ones; each should be meaningful enough to justify the effort and governance required. After the workshop, circulate the briefs for feedback and refinement.
Step 4: Do a Reality Check
Now comes the backroom work. For each opportunity, run honest technical due diligence—check data quality, availability, integration complexity, and security. Evaluate privacy and compliance early, as regulatory reviews or DPIAs can delay delivery. Build realistic cost models that include integration, maintenance, and governance overhead, not just AI platform costs. Most teams underestimate these, so address them upfront.
Step 5: Prioritisation and Selection
Bring the same group from Step 3 together and score each opportunity using a simple 2×2 matrix — value to the business (y-axis) versus ease of implementation (x-axis). Rankings can be relative but must reflect real insights from earlier steps. Assess impact across the whole organisation, not just one team, and consider cultural barriers as well as technical ones. Choose one opportunity to pilot and one to incubate, avoiding multiple parallel pilots that dilute focus. Finally, document why others weren’t selected — it keeps future decisions grounded.
Step 6: Build the Minimum Viable Product
Your “thin slice” must account for data dependencies. A chatbot MVP shouldn’t aim to answer 20 questions—it should test whether your data is usable, your AI understands domain language, and your tone stays consistent. Start with a simple, high-quality data pipeline—maybe 50 curated Q&As instead of full automation.
AI MVPs can’t be faked with humans like traditional ones. Focus on minimal scope: one document type, one department, or one product category. Begin simple—prove the concept before scaling.
The hardest lessons come from integration, not accuracy. Can the AI access data, fit into workflows, and drive adoption? A basic rule-based system that works in context beats a sophisticated model in isolation. Use sandbox or read-only environments to stay safe, and build monitoring and guardrails early. Track metrics like quality, latency, cost, and user satisfaction from day one.
Test adoption risk first, not technical depth. For fraud detection, test analyst trust; for content tools, test output quality. Avoid endless tweaks—ship early, learn fast. Real feedback on a small working system beats perfection delayed.
Step 7: Test, Learn, Iterate
With your MVP live, run it on historical data or alongside existing processes to validate assumptions against real results. Gather feedback from a small group of friendly users to complement your metrics with real insights. Follow the Lean Startup loop – fix only the top three issues, avoiding scope creep. Document incidents, edge cases, and feedback carefully; this learning will guide your go/no-go decision.
Step 8: Evaluate Results Against Hypotheses
Summarise your findings in a clear evaluation document linked to your original goals—compare results to baselines, review costs, note risks, and capture user feedback. Watch for warning signs that signal stopping over scaling: rising costs, mounting data issues, poor adoption, stagnant accuracy, or complex integrations. Use innovation accounting to judge progress and decide whether to scale, pivot, or pause. Be transparent about what worked, what failed, and what remains uncertain.
Step 9: Pivot, Persevere or Scale Decision
If your testing validates your core hypotheses, establish runtime service level objectives, assign clear ownership, set cost caps, and define support arrangements for scaling. Make these commitments visible to all stakeholders.
If pivoting, clearly document what you've learned and how it informs your new direction. A pivot isn't a failure—it's a structured course correction based on validated learning.
If stopping, archive your technical assets properly and communicate the decision transparently. As Ries emphasises, failed experiments that generate clear learnings are valuable investments, not wasted effort.
Communicate your decision and rationale to all stakeholders. Clear communication maintains trust and supports future initiatives regardless of which path you choose.
Sprint Outcomes
At the end of your sprint, you'll have:
Business Model Snapshot: Current-state view using the nine building blocks framework
Initiative Briefs: 3-5 one-page opportunities with clear business justification
Prioritisation Analysis: Scored evaluation with transparent decision criteria
Experiment Plan: Detailed 30-day test plan for your selected opportunity
Decision Package: Evidence-based recommendation with clear next steps
Pilot Runbook: Operational guidance for running your pilot successfully
3. Your Step-by-Step Guide
Step 1: Foundation and Goal Setting
Begin by defining exactly what your business aims to achieve — growth, margin improvement, risk reduction, or a better customer experience. Be specific with targets and timeframes. This isn’t about AI yet; it’s about understanding what your business delivers and what success looks like. Focus on outcome-based metrics that matter, such as “Reduce customer service cycle time by 30%” rather than “achieve 95% model accuracy.” Set thresholds for what counts as real improvement.
Next, introduce the Business Model Canvas to map how your organisation creates, delivers, and captures value. You and your team will use it to identify where AI can make a meaningful difference and to assess the impact of any changes. This shared structure becomes the foundation for every AI conversation that follows.
Step 2: Current State Reality Check
Now, make sure your business model canvas is complete. This means gathering real data on how your business operates today. Have someone walk the floors, document key processes, customer journeys, and decision points — including timing, handoffs, and failure modes. Add metrics wherever possible. Understanding how customers actually interact with you reveals where automation and business model improvements may lie.
Next, create an inventory of your systems, data sources, and integrations. Be brutally honest about data quality, access restrictions, and technical debt — most AI projects fail because these issues surface too late. Finally, document your cost structure, revenue streams, and pain points — including customer friction, inefficiencies, and regulatory or technical constraints. This becomes your foundation for evaluating where AI can drive real, sustainable value.
Step 3: AI Opportunity Discovery
Work through each of the nine canvas areas systematically, asking how AI might change the dynamics in each area. Whilst its tempting to leave this to one person to do, many minds are better than one. This works best when a diverse range of people from across your organisation get together in a room. The person who did the legwork in understanding your business in Step 2 should present, the factual accuracy should be agreed, and then start to brainstorm ideas.
That can be the hard bit, and the format of the workshop needs to be carefully managed to help this. Understanding who you have in the room, and how to get the most out of them is critical. There are plenty of good techniques that you can use to effectively brainstorm.
Consider inviting key customers or suppliers to join for an outside perspective.
For each opportunity, create a one-page brief focused on the business problem first, then how AI solves it – clearly defining what it will and won’t do. Aim for three to five substantial opportunities rather than a long list of minor ones; each should be meaningful enough to justify the effort and governance required. After the workshop, circulate the briefs for feedback and refinement.
Step 4: Do a Reality Check
Now comes the backroom work. For each opportunity, run honest technical due diligence—check data quality, availability, integration complexity, and security. Evaluate privacy and compliance early, as regulatory reviews or DPIAs can delay delivery. Build realistic cost models that include integration, maintenance, and governance overhead, not just AI platform costs. Most teams underestimate these, so address them upfront.
Step 5: Prioritisation and Selection
Bring the same group from Step 3 together and score each opportunity using a simple 2×2 matrix — value to the business (y-axis) versus ease of implementation (x-axis). Rankings can be relative but must reflect real insights from earlier steps. Assess impact across the whole organisation, not just one team, and consider cultural barriers as well as technical ones. Choose one opportunity to pilot and one to incubate, avoiding multiple parallel pilots that dilute focus. Finally, document why others weren’t selected — it keeps future decisions grounded.
Step 6: Build the Minimum Viable Product
Your “thin slice” must account for data dependencies. A chatbot MVP shouldn’t aim to answer 20 questions—it should test whether your data is usable, your AI understands domain language, and your tone stays consistent. Start with a simple, high-quality data pipeline—maybe 50 curated Q&As instead of full automation.
AI MVPs can’t be faked with humans like traditional ones. Focus on minimal scope: one document type, one department, or one product category. Begin simple—prove the concept before scaling.
The hardest lessons come from integration, not accuracy. Can the AI access data, fit into workflows, and drive adoption? A basic rule-based system that works in context beats a sophisticated model in isolation. Use sandbox or read-only environments to stay safe, and build monitoring and guardrails early. Track metrics like quality, latency, cost, and user satisfaction from day one.
Test adoption risk first, not technical depth. For fraud detection, test analyst trust; for content tools, test output quality. Avoid endless tweaks—ship early, learn fast. Real feedback on a small working system beats perfection delayed.
Step 7: Test, Learn, Iterate
With your MVP live, run it on historical data or alongside existing processes to validate assumptions against real results. Gather feedback from a small group of friendly users to complement your metrics with real insights. Follow the Lean Startup loop – fix only the top three issues, avoiding scope creep. Document incidents, edge cases, and feedback carefully; this learning will guide your go/no-go decision.
Step 8: Evaluate Results Against Hypotheses
Summarise your findings in a clear evaluation document linked to your original goals—compare results to baselines, review costs, note risks, and capture user feedback. Watch for warning signs that signal stopping over scaling: rising costs, mounting data issues, poor adoption, stagnant accuracy, or complex integrations. Use innovation accounting to judge progress and decide whether to scale, pivot, or pause. Be transparent about what worked, what failed, and what remains uncertain.
Step 9: Pivot, Persevere or Scale Decision
If your testing validates your core hypotheses, establish runtime service level objectives, assign clear ownership, set cost caps, and define support arrangements for scaling. Make these commitments visible to all stakeholders.
If pivoting, clearly document what you've learned and how it informs your new direction. A pivot isn't a failure—it's a structured course correction based on validated learning.
If stopping, archive your technical assets properly and communicate the decision transparently. As Ries emphasises, failed experiments that generate clear learnings are valuable investments, not wasted effort.
Communicate your decision and rationale to all stakeholders. Clear communication maintains trust and supports future initiatives regardless of which path you choose.
Sprint Outcomes
At the end of your sprint, you'll have:
Business Model Snapshot: Current-state view using the nine building blocks framework
Initiative Briefs: 3-5 one-page opportunities with clear business justification
Prioritisation Analysis: Scored evaluation with transparent decision criteria
Experiment Plan: Detailed 30-day test plan for your selected opportunity
Decision Package: Evidence-based recommendation with clear next steps
Pilot Runbook: Operational guidance for running your pilot successfully
3. Your Step-by-Step Guide
Step 1: Foundation and Goal Setting
Begin by defining exactly what your business aims to achieve — growth, margin improvement, risk reduction, or a better customer experience. Be specific with targets and timeframes. This isn’t about AI yet; it’s about understanding what your business delivers and what success looks like. Focus on outcome-based metrics that matter, such as “Reduce customer service cycle time by 30%” rather than “achieve 95% model accuracy.” Set thresholds for what counts as real improvement.
Next, introduce the Business Model Canvas to map how your organisation creates, delivers, and captures value. You and your team will use it to identify where AI can make a meaningful difference and to assess the impact of any changes. This shared structure becomes the foundation for every AI conversation that follows.
Step 2: Current State Reality Check
Now, make sure your business model canvas is complete. This means gathering real data on how your business operates today. Have someone walk the floors, document key processes, customer journeys, and decision points — including timing, handoffs, and failure modes. Add metrics wherever possible. Understanding how customers actually interact with you reveals where automation and business model improvements may lie.
Next, create an inventory of your systems, data sources, and integrations. Be brutally honest about data quality, access restrictions, and technical debt — most AI projects fail because these issues surface too late. Finally, document your cost structure, revenue streams, and pain points — including customer friction, inefficiencies, and regulatory or technical constraints. This becomes your foundation for evaluating where AI can drive real, sustainable value.
Step 3: AI Opportunity Discovery
Work through each of the nine canvas areas systematically, asking how AI might change the dynamics in each area. Whilst its tempting to leave this to one person to do, many minds are better than one. This works best when a diverse range of people from across your organisation get together in a room. The person who did the legwork in understanding your business in Step 2 should present, the factual accuracy should be agreed, and then start to brainstorm ideas.
That can be the hard bit, and the format of the workshop needs to be carefully managed to help this. Understanding who you have in the room, and how to get the most out of them is critical. There are plenty of good techniques that you can use to effectively brainstorm.
Consider inviting key customers or suppliers to join for an outside perspective.
For each opportunity, create a one-page brief focused on the business problem first, then how AI solves it – clearly defining what it will and won’t do. Aim for three to five substantial opportunities rather than a long list of minor ones; each should be meaningful enough to justify the effort and governance required. After the workshop, circulate the briefs for feedback and refinement.
Step 4: Do a Reality Check
Now comes the backroom work. For each opportunity, run honest technical due diligence—check data quality, availability, integration complexity, and security. Evaluate privacy and compliance early, as regulatory reviews or DPIAs can delay delivery. Build realistic cost models that include integration, maintenance, and governance overhead, not just AI platform costs. Most teams underestimate these, so address them upfront.
Step 5: Prioritisation and Selection
Bring the same group from Step 3 together and score each opportunity using a simple 2×2 matrix — value to the business (y-axis) versus ease of implementation (x-axis). Rankings can be relative but must reflect real insights from earlier steps. Assess impact across the whole organisation, not just one team, and consider cultural barriers as well as technical ones. Choose one opportunity to pilot and one to incubate, avoiding multiple parallel pilots that dilute focus. Finally, document why others weren’t selected — it keeps future decisions grounded.
Step 6: Build the Minimum Viable Product
Your “thin slice” must account for data dependencies. A chatbot MVP shouldn’t aim to answer 20 questions—it should test whether your data is usable, your AI understands domain language, and your tone stays consistent. Start with a simple, high-quality data pipeline—maybe 50 curated Q&As instead of full automation.
AI MVPs can’t be faked with humans like traditional ones. Focus on minimal scope: one document type, one department, or one product category. Begin simple—prove the concept before scaling.
The hardest lessons come from integration, not accuracy. Can the AI access data, fit into workflows, and drive adoption? A basic rule-based system that works in context beats a sophisticated model in isolation. Use sandbox or read-only environments to stay safe, and build monitoring and guardrails early. Track metrics like quality, latency, cost, and user satisfaction from day one.
Test adoption risk first, not technical depth. For fraud detection, test analyst trust; for content tools, test output quality. Avoid endless tweaks—ship early, learn fast. Real feedback on a small working system beats perfection delayed.
Step 7: Test, Learn, Iterate
With your MVP live, run it on historical data or alongside existing processes to validate assumptions against real results. Gather feedback from a small group of friendly users to complement your metrics with real insights. Follow the Lean Startup loop – fix only the top three issues, avoiding scope creep. Document incidents, edge cases, and feedback carefully; this learning will guide your go/no-go decision.
Step 8: Evaluate Results Against Hypotheses
Summarise your findings in a clear evaluation document linked to your original goals—compare results to baselines, review costs, note risks, and capture user feedback. Watch for warning signs that signal stopping over scaling: rising costs, mounting data issues, poor adoption, stagnant accuracy, or complex integrations. Use innovation accounting to judge progress and decide whether to scale, pivot, or pause. Be transparent about what worked, what failed, and what remains uncertain.
Step 9: Pivot, Persevere or Scale Decision
If your testing validates your core hypotheses, establish runtime service level objectives, assign clear ownership, set cost caps, and define support arrangements for scaling. Make these commitments visible to all stakeholders.
If pivoting, clearly document what you've learned and how it informs your new direction. A pivot isn't a failure—it's a structured course correction based on validated learning.
If stopping, archive your technical assets properly and communicate the decision transparently. As Ries emphasises, failed experiments that generate clear learnings are valuable investments, not wasted effort.
Communicate your decision and rationale to all stakeholders. Clear communication maintains trust and supports future initiatives regardless of which path you choose.
Sprint Outcomes
At the end of your sprint, you'll have:
Business Model Snapshot: Current-state view using the nine building blocks framework
Initiative Briefs: 3-5 one-page opportunities with clear business justification
Prioritisation Analysis: Scored evaluation with transparent decision criteria
Experiment Plan: Detailed 30-day test plan for your selected opportunity
Decision Package: Evidence-based recommendation with clear next steps
Pilot Runbook: Operational guidance for running your pilot successfully
4. Know Your Business Model
Before exploring AI applications, you need crystal-clear understanding of how your organisation creates, delivers, and captures value today. The business model canvas, developed by Strategyzer, provides the systematic approach you need.
The Business Model Canvas is based around nine areas, we call these the Nine Building Blocks.
Nine Building Blocks Decoded
Think of these as the core building blocks of your business model—the blueprint showing how everything connects. Understanding these relationships helps you anticipate how AI-driven changes in one area affect others.
Customer Segments are distinct groups with unique needs and behaviours. Each responds differently to AI enhancements depending on their goals and expectations.
Value Propositions define the outcomes you deliver, not just the features—AI supports these results rather than replacing them.
Channels cover every point of customer interaction. AI can personalise, automate, and streamline these touchpoints but must align with audience preferences.
Customer Relationships shape how you attract and retain clients; AI enhances them with proactive, personalised service while keeping human connection.
Revenue Streams explain how value turns into income. AI can unlock new pricing models or improve existing ones.
Key Resources include data, talent, and technology—the assets essential to deliver on your promise.
Key Activities are what your organisation must excel at; AI makes these faster, smarter, or more scalable.
Key Partners bring specialised capabilities or data that make AI viable.
Cost Structure captures how you spend to create value – AI shifts this toward compute and data management rather than labour.
How AI Transforms Business Models
AI doesn't just automate existing processes—it changes fundamental economics and possibilities within each building block.
Scarcity Shifts: AI changes what's scarce and abundant. Tasks requiring extensive human expertise become scalable through AI, while human skills like creativity, empathy, and strategic thinking become more valuable. This can make previously uneconomic segments viable or enable entirely new value propositions.
Economic Changes: AI introduces different cost curves and revenue possibilities. Instead of linear scaling costs, you get high upfront investment with low marginal costs. Instead of fixed pricing, you can enable usage-based or outcome-based models. These changes alter competitive dynamics fundamentally.
New Risk Categories: AI introduces unprecedented risks including model bias, data privacy violations, algorithmic transparency requirements, and vendor concentration. These don't just affect compliance—they impact customer trust and competitive position.
Speed and Scale Opportunities: AI compresses timeframes and expands reach in ways that change competition entirely. Processes that took days happen in seconds. Services requiring local expertise become globally available instantly.
4. Know Your Business Model
Before exploring AI applications, you need crystal-clear understanding of how your organisation creates, delivers, and captures value today. The business model canvas, developed by Strategyzer, provides the systematic approach you need.
The Business Model Canvas is based around nine areas, we call these the Nine Building Blocks.
Nine Building Blocks Decoded
Think of these as the core building blocks of your business model—the blueprint showing how everything connects. Understanding these relationships helps you anticipate how AI-driven changes in one area affect others.
Customer Segments are distinct groups with unique needs and behaviours. Each responds differently to AI enhancements depending on their goals and expectations.
Value Propositions define the outcomes you deliver, not just the features—AI supports these results rather than replacing them.
Channels cover every point of customer interaction. AI can personalise, automate, and streamline these touchpoints but must align with audience preferences.
Customer Relationships shape how you attract and retain clients; AI enhances them with proactive, personalised service while keeping human connection.
Revenue Streams explain how value turns into income. AI can unlock new pricing models or improve existing ones.
Key Resources include data, talent, and technology—the assets essential to deliver on your promise.
Key Activities are what your organisation must excel at; AI makes these faster, smarter, or more scalable.
Key Partners bring specialised capabilities or data that make AI viable.
Cost Structure captures how you spend to create value – AI shifts this toward compute and data management rather than labour.
How AI Transforms Business Models
AI doesn't just automate existing processes—it changes fundamental economics and possibilities within each building block.
Scarcity Shifts: AI changes what's scarce and abundant. Tasks requiring extensive human expertise become scalable through AI, while human skills like creativity, empathy, and strategic thinking become more valuable. This can make previously uneconomic segments viable or enable entirely new value propositions.
Economic Changes: AI introduces different cost curves and revenue possibilities. Instead of linear scaling costs, you get high upfront investment with low marginal costs. Instead of fixed pricing, you can enable usage-based or outcome-based models. These changes alter competitive dynamics fundamentally.
New Risk Categories: AI introduces unprecedented risks including model bias, data privacy violations, algorithmic transparency requirements, and vendor concentration. These don't just affect compliance—they impact customer trust and competitive position.
Speed and Scale Opportunities: AI compresses timeframes and expands reach in ways that change competition entirely. Processes that took days happen in seconds. Services requiring local expertise become globally available instantly.
4. Know Your Business Model
Before exploring AI applications, you need crystal-clear understanding of how your organisation creates, delivers, and captures value today. The business model canvas, developed by Strategyzer, provides the systematic approach you need.
The Business Model Canvas is based around nine areas, we call these the Nine Building Blocks.
Nine Building Blocks Decoded
Think of these as the core building blocks of your business model—the blueprint showing how everything connects. Understanding these relationships helps you anticipate how AI-driven changes in one area affect others.
Customer Segments are distinct groups with unique needs and behaviours. Each responds differently to AI enhancements depending on their goals and expectations.
Value Propositions define the outcomes you deliver, not just the features—AI supports these results rather than replacing them.
Channels cover every point of customer interaction. AI can personalise, automate, and streamline these touchpoints but must align with audience preferences.
Customer Relationships shape how you attract and retain clients; AI enhances them with proactive, personalised service while keeping human connection.
Revenue Streams explain how value turns into income. AI can unlock new pricing models or improve existing ones.
Key Resources include data, talent, and technology—the assets essential to deliver on your promise.
Key Activities are what your organisation must excel at; AI makes these faster, smarter, or more scalable.
Key Partners bring specialised capabilities or data that make AI viable.
Cost Structure captures how you spend to create value – AI shifts this toward compute and data management rather than labour.
How AI Transforms Business Models
AI doesn't just automate existing processes—it changes fundamental economics and possibilities within each building block.
Scarcity Shifts: AI changes what's scarce and abundant. Tasks requiring extensive human expertise become scalable through AI, while human skills like creativity, empathy, and strategic thinking become more valuable. This can make previously uneconomic segments viable or enable entirely new value propositions.
Economic Changes: AI introduces different cost curves and revenue possibilities. Instead of linear scaling costs, you get high upfront investment with low marginal costs. Instead of fixed pricing, you can enable usage-based or outcome-based models. These changes alter competitive dynamics fundamentally.
New Risk Categories: AI introduces unprecedented risks including model bias, data privacy violations, algorithmic transparency requirements, and vendor concentration. These don't just affect compliance—they impact customer trust and competitive position.
Speed and Scale Opportunities: AI compresses timeframes and expands reach in ways that change competition entirely. Processes that took days happen in seconds. Services requiring local expertise become globally available instantly.
5. AI Opportunity Analysis
Now you can systematically explore how AI might transform each business model aspect. The key is thinking beyond automation. The most valuable opportunities often involve entirely new capabilities or fundamental business dynamic changes.
Within your team you should establish ground rules for decision making, here are some we like:
Facts First: Base every discussion on real data, not assumptions. Mark guesses in italics so they can be tested later.
Measure Everything: Link every opportunity, success, and risk to a measurable outcome. Define what proof would confirm or disprove it.
Stay Focused: Clearly state what each initiative includes—and excludes. Scope creep kills more AI projects than tech issues.
Examples of where AI can help?
Let's explore some practical AI applications. For each area, we've identified high-impact opportunities that can deliver measurable value with reasonable implementation complexity. These examples serve as starting points for your own opportunity analysis.
Customer Segments: Precision and Personalisation
Traditional segmentation relies on demographics, past behaviour, or manual criteria. AI enables dynamic, behaviour-based segmentation revealing hidden patterns and "segment of one" personalisation.
Discovery Questions
Which segments generate highest value but receive inconsistent service quality?
Where do you see concentrated churn rates and what early signals might predict departures?
Which customer needs are visible in your data but currently uneconomic to serve?
Quick Wins
Churn Prediction and Prevention: Build models identifying at-risk customers based on usage patterns, support interactions, and engagement metrics
Lookalike Customer Discovery: Use AI to analyse your best customers and identify similar prospects
Dynamic Retention Offers: Automatically trigger personalised retention offers based on individual customer value and churn probability
30-Day Experiment Framework
Select a customer segment with clear churn patterns and sufficient historical data. Build a simple churn prediction model using readily available sources. Arm one customer success team with predictions and response playbooks. Measure churn rate changes versus a matched control group.
Value Propositions: Enhancing Customer Outcomes
AI transforms value propositions by changing speed, accuracy, personalisation, or scope of outcomes you deliver. Identify where AI enables previously impossible or impractical customer outcomes.
Quick Wins
Intelligent Form Completion: Pre-populate forms, validate information real-time, and guide customers through complex processes
Document Analysis and Summarisation: Automatically extract key information from complex documents and present relevant summaries
Quality Assurance with Confidence Scoring: Add AI-powered quality checks flagging potential issues with confidence scores
Channels: Intelligent Customer Acquisition and Service
AI transforms how customers discover, evaluate, purchase, and receive value from your organisation.
Quick Wins
FAQ Automation with Intelligent Routing: Build retrieval systems answering common questions using existing documentation
Next-Best-Action Recommendations: Use AI suggesting most relevant content, offers, or actions based on prospect behaviour
Intelligent Lead Scoring: Automatically score and route leads based on conversion likelihood and potential value
Customer Relationships: Proactive and Personalised Service
AI transforms relationships by enabling proactive support, personalised communications, and intelligent service delivery anticipating needs rather than just responding to problems.
Quick Wins
Agent Assistance Tools: Provide representatives with AI-generated customer history summaries and suggested responses
Automated Case Routing: Use AI analysing incoming requests and routing to most appropriate teams
Sentiment Monitoring: Automatically detect negative sentiment and escalate for immediate attention
5. AI Opportunity Analysis
Now you can systematically explore how AI might transform each business model aspect. The key is thinking beyond automation. The most valuable opportunities often involve entirely new capabilities or fundamental business dynamic changes.
Within your team you should establish ground rules for decision making, here are some we like:
Facts First: Base every discussion on real data, not assumptions. Mark guesses in italics so they can be tested later.
Measure Everything: Link every opportunity, success, and risk to a measurable outcome. Define what proof would confirm or disprove it.
Stay Focused: Clearly state what each initiative includes—and excludes. Scope creep kills more AI projects than tech issues.
Examples of where AI can help?
Let's explore some practical AI applications. For each area, we've identified high-impact opportunities that can deliver measurable value with reasonable implementation complexity. These examples serve as starting points for your own opportunity analysis.
Customer Segments: Precision and Personalisation
Traditional segmentation relies on demographics, past behaviour, or manual criteria. AI enables dynamic, behaviour-based segmentation revealing hidden patterns and "segment of one" personalisation.
Discovery Questions
Which segments generate highest value but receive inconsistent service quality?
Where do you see concentrated churn rates and what early signals might predict departures?
Which customer needs are visible in your data but currently uneconomic to serve?
Quick Wins
Churn Prediction and Prevention: Build models identifying at-risk customers based on usage patterns, support interactions, and engagement metrics
Lookalike Customer Discovery: Use AI to analyse your best customers and identify similar prospects
Dynamic Retention Offers: Automatically trigger personalised retention offers based on individual customer value and churn probability
30-Day Experiment Framework
Select a customer segment with clear churn patterns and sufficient historical data. Build a simple churn prediction model using readily available sources. Arm one customer success team with predictions and response playbooks. Measure churn rate changes versus a matched control group.
Value Propositions: Enhancing Customer Outcomes
AI transforms value propositions by changing speed, accuracy, personalisation, or scope of outcomes you deliver. Identify where AI enables previously impossible or impractical customer outcomes.
Quick Wins
Intelligent Form Completion: Pre-populate forms, validate information real-time, and guide customers through complex processes
Document Analysis and Summarisation: Automatically extract key information from complex documents and present relevant summaries
Quality Assurance with Confidence Scoring: Add AI-powered quality checks flagging potential issues with confidence scores
Channels: Intelligent Customer Acquisition and Service
AI transforms how customers discover, evaluate, purchase, and receive value from your organisation.
Quick Wins
FAQ Automation with Intelligent Routing: Build retrieval systems answering common questions using existing documentation
Next-Best-Action Recommendations: Use AI suggesting most relevant content, offers, or actions based on prospect behaviour
Intelligent Lead Scoring: Automatically score and route leads based on conversion likelihood and potential value
Customer Relationships: Proactive and Personalised Service
AI transforms relationships by enabling proactive support, personalised communications, and intelligent service delivery anticipating needs rather than just responding to problems.
Quick Wins
Agent Assistance Tools: Provide representatives with AI-generated customer history summaries and suggested responses
Automated Case Routing: Use AI analysing incoming requests and routing to most appropriate teams
Sentiment Monitoring: Automatically detect negative sentiment and escalate for immediate attention
5. AI Opportunity Analysis
Now you can systematically explore how AI might transform each business model aspect. The key is thinking beyond automation. The most valuable opportunities often involve entirely new capabilities or fundamental business dynamic changes.
Within your team you should establish ground rules for decision making, here are some we like:
Facts First: Base every discussion on real data, not assumptions. Mark guesses in italics so they can be tested later.
Measure Everything: Link every opportunity, success, and risk to a measurable outcome. Define what proof would confirm or disprove it.
Stay Focused: Clearly state what each initiative includes—and excludes. Scope creep kills more AI projects than tech issues.
Examples of where AI can help?
Let's explore some practical AI applications. For each area, we've identified high-impact opportunities that can deliver measurable value with reasonable implementation complexity. These examples serve as starting points for your own opportunity analysis.
Customer Segments: Precision and Personalisation
Traditional segmentation relies on demographics, past behaviour, or manual criteria. AI enables dynamic, behaviour-based segmentation revealing hidden patterns and "segment of one" personalisation.
Discovery Questions
Which segments generate highest value but receive inconsistent service quality?
Where do you see concentrated churn rates and what early signals might predict departures?
Which customer needs are visible in your data but currently uneconomic to serve?
Quick Wins
Churn Prediction and Prevention: Build models identifying at-risk customers based on usage patterns, support interactions, and engagement metrics
Lookalike Customer Discovery: Use AI to analyse your best customers and identify similar prospects
Dynamic Retention Offers: Automatically trigger personalised retention offers based on individual customer value and churn probability
30-Day Experiment Framework
Select a customer segment with clear churn patterns and sufficient historical data. Build a simple churn prediction model using readily available sources. Arm one customer success team with predictions and response playbooks. Measure churn rate changes versus a matched control group.
Value Propositions: Enhancing Customer Outcomes
AI transforms value propositions by changing speed, accuracy, personalisation, or scope of outcomes you deliver. Identify where AI enables previously impossible or impractical customer outcomes.
Quick Wins
Intelligent Form Completion: Pre-populate forms, validate information real-time, and guide customers through complex processes
Document Analysis and Summarisation: Automatically extract key information from complex documents and present relevant summaries
Quality Assurance with Confidence Scoring: Add AI-powered quality checks flagging potential issues with confidence scores
Channels: Intelligent Customer Acquisition and Service
AI transforms how customers discover, evaluate, purchase, and receive value from your organisation.
Quick Wins
FAQ Automation with Intelligent Routing: Build retrieval systems answering common questions using existing documentation
Next-Best-Action Recommendations: Use AI suggesting most relevant content, offers, or actions based on prospect behaviour
Intelligent Lead Scoring: Automatically score and route leads based on conversion likelihood and potential value
Customer Relationships: Proactive and Personalised Service
AI transforms relationships by enabling proactive support, personalised communications, and intelligent service delivery anticipating needs rather than just responding to problems.
Quick Wins
Agent Assistance Tools: Provide representatives with AI-generated customer history summaries and suggested responses
Automated Case Routing: Use AI analysing incoming requests and routing to most appropriate teams
Sentiment Monitoring: Automatically detect negative sentiment and escalate for immediate attention
6. Making Hard Choices
With multiple AI opportunities identified, you need a systematic approach that cuts through complexity and delivers clear decisions. The most effective prioritisation method is often the simplest one – a straightforward 2x2 matrix that creates objective comparison across very different types of opportunities.
The 2x2 Matrix Approach: Simple but Powerful
If possible, bring back the same team from your Step 3 workshop—they already share context and understand the business realities behind each idea, which keeps evaluation consistent. Use a simple 2×2 matrix to prioritise opportunities: ease of implementation on the X-axis and value to the entire business on the Y-axis. You don’t need complex financial models; relative rankings grounded in what you learned earlier are enough to guide clear, aligned decisions.
Common Prioritisation Mistakes to Avoid
Mistake #1: Team-Level Thinking
What appears high-impact for one team may deliver minimal value to the broader organisation. A customer service AI that reduces response time by 50% seems transformative to the support team, but if it only affects 2% of customer interactions, the organisation-wide impact is limited.
Always ask: "How does this create value for the entire business, not just this department?"
Mistake #2: Wishful Thinking About Implementation
Teams consistently underestimate implementation difficulty – especially cultural barriers. Technical feasibility is only part of the picture; you must also assess data integration complexity, system changes, cultural and process shifts, and any regulatory or compliance hurdles. This is why the reality check from Step 4 is so critical – your ease-of-implementation assessment must be grounded in facts, not optimism.
The Cultural Barrier Reality Check
Don't underestimate the cultural barriers that need to be overcome to make change happen. Technical solutions often fail due to human resistance rather than technical limitations.
Ask tough questions about each opportunity:
Will people actually use this, or will they find workarounds?
What behaviours must change, and how realistic is that change?
Who might resist this change, and do we have plans to address their concerns?
Are the required process changes compatible with existing workflows?
Cultural barriers often move opportunities from "easy to implement" to "complex change management required."
Making the Selection Decision
Select Two Opportunities Maximum
Choose one opportunity to pilot immediately and one to incubate for future consideration. This disciplined approach ensures adequate resources and attention for generating clear learnings.
Resist the Multiple Pilot Temptation
The instinct to pursue several pilots simultaneously almost always leads to under-resourced initiatives that fail to generate clear learnings. You end up with multiple inconclusive experiments rather than one definitive success or failure.
Start small and learn. One well-executed pilot teaches you more about AI implementation than three under-resourced experiments.
Documentation for Future Context
Record Your Reasoning
Document—or at least clearly remember—why you're not pursuing the other opportunities. This serves two critical purposes:
Maintains Focus: Clear reasoning helps resist scope creep and feature requests that pull attention from your chosen pilot
Provides Context: Future prioritisation decisions benefit from understanding previous logic, especially when conditions change
Sample Documentation Format:
For each opportunity not selected, capture:
Why it scored lower: Specific factors that reduced its priority
What would need to change: Conditions that might make it viable later
Timeline considerations: When it might be appropriate to revisit
Matrix Positioning Guide
Opportunities that offer high value and easy implementation are your best pilot candidates—they promise meaningful business impact without overwhelming complexity. Start here to demonstrate quick, visible results.
Those with high value but hard implementation should be incubated for later. Their potential payoff warrants investment once your organisation strengthens its technical and operational capabilities.
Meanwhile, low-value but easy-to-implement ideas can serve as small wins to build confidence, but don’t let them distract from high-impact work. Anything that’s both low value and hard to implement should be parked for now—save your resources for opportunities that truly move the needle.
Making It Practical
Scoring Approach: Use a simple 1-5 scale for each axis, but focus on relative positioning rather than precise numerical scores. The goal is clear differentiation between opportunities, not mathematical precision.
Decision Criteria: The highest combined score wins, but use your judgment about quadrant positioning. A 4,4 opportunity (high value, relatively easy) might be preferable to a 5,3 opportunity (very high value, moderate difficulty) depending on your organisation's current capability and risk tolerance.
Timeline Pressure: If you're under pressure to move quickly, bias toward ease of implementation. Success with a moderately valuable but achievable pilot builds credibility for tackling harder, higher-value opportunities later.
6. Making Hard Choices
With multiple AI opportunities identified, you need a systematic approach that cuts through complexity and delivers clear decisions. The most effective prioritisation method is often the simplest one – a straightforward 2x2 matrix that creates objective comparison across very different types of opportunities.
The 2x2 Matrix Approach: Simple but Powerful
If possible, bring back the same team from your Step 3 workshop—they already share context and understand the business realities behind each idea, which keeps evaluation consistent. Use a simple 2×2 matrix to prioritise opportunities: ease of implementation on the X-axis and value to the entire business on the Y-axis. You don’t need complex financial models; relative rankings grounded in what you learned earlier are enough to guide clear, aligned decisions.
Common Prioritisation Mistakes to Avoid
Mistake #1: Team-Level Thinking
What appears high-impact for one team may deliver minimal value to the broader organisation. A customer service AI that reduces response time by 50% seems transformative to the support team, but if it only affects 2% of customer interactions, the organisation-wide impact is limited.
Always ask: "How does this create value for the entire business, not just this department?"
Mistake #2: Wishful Thinking About Implementation
Teams consistently underestimate implementation difficulty – especially cultural barriers. Technical feasibility is only part of the picture; you must also assess data integration complexity, system changes, cultural and process shifts, and any regulatory or compliance hurdles. This is why the reality check from Step 4 is so critical – your ease-of-implementation assessment must be grounded in facts, not optimism.
The Cultural Barrier Reality Check
Don't underestimate the cultural barriers that need to be overcome to make change happen. Technical solutions often fail due to human resistance rather than technical limitations.
Ask tough questions about each opportunity:
Will people actually use this, or will they find workarounds?
What behaviours must change, and how realistic is that change?
Who might resist this change, and do we have plans to address their concerns?
Are the required process changes compatible with existing workflows?
Cultural barriers often move opportunities from "easy to implement" to "complex change management required."
Making the Selection Decision
Select Two Opportunities Maximum
Choose one opportunity to pilot immediately and one to incubate for future consideration. This disciplined approach ensures adequate resources and attention for generating clear learnings.
Resist the Multiple Pilot Temptation
The instinct to pursue several pilots simultaneously almost always leads to under-resourced initiatives that fail to generate clear learnings. You end up with multiple inconclusive experiments rather than one definitive success or failure.
Start small and learn. One well-executed pilot teaches you more about AI implementation than three under-resourced experiments.
Documentation for Future Context
Record Your Reasoning
Document—or at least clearly remember—why you're not pursuing the other opportunities. This serves two critical purposes:
Maintains Focus: Clear reasoning helps resist scope creep and feature requests that pull attention from your chosen pilot
Provides Context: Future prioritisation decisions benefit from understanding previous logic, especially when conditions change
Sample Documentation Format:
For each opportunity not selected, capture:
Why it scored lower: Specific factors that reduced its priority
What would need to change: Conditions that might make it viable later
Timeline considerations: When it might be appropriate to revisit
Matrix Positioning Guide
Opportunities that offer high value and easy implementation are your best pilot candidates—they promise meaningful business impact without overwhelming complexity. Start here to demonstrate quick, visible results.
Those with high value but hard implementation should be incubated for later. Their potential payoff warrants investment once your organisation strengthens its technical and operational capabilities.
Meanwhile, low-value but easy-to-implement ideas can serve as small wins to build confidence, but don’t let them distract from high-impact work. Anything that’s both low value and hard to implement should be parked for now—save your resources for opportunities that truly move the needle.
Making It Practical
Scoring Approach: Use a simple 1-5 scale for each axis, but focus on relative positioning rather than precise numerical scores. The goal is clear differentiation between opportunities, not mathematical precision.
Decision Criteria: The highest combined score wins, but use your judgment about quadrant positioning. A 4,4 opportunity (high value, relatively easy) might be preferable to a 5,3 opportunity (very high value, moderate difficulty) depending on your organisation's current capability and risk tolerance.
Timeline Pressure: If you're under pressure to move quickly, bias toward ease of implementation. Success with a moderately valuable but achievable pilot builds credibility for tackling harder, higher-value opportunities later.
6. Making Hard Choices
With multiple AI opportunities identified, you need a systematic approach that cuts through complexity and delivers clear decisions. The most effective prioritisation method is often the simplest one – a straightforward 2x2 matrix that creates objective comparison across very different types of opportunities.
The 2x2 Matrix Approach: Simple but Powerful
If possible, bring back the same team from your Step 3 workshop—they already share context and understand the business realities behind each idea, which keeps evaluation consistent. Use a simple 2×2 matrix to prioritise opportunities: ease of implementation on the X-axis and value to the entire business on the Y-axis. You don’t need complex financial models; relative rankings grounded in what you learned earlier are enough to guide clear, aligned decisions.
Common Prioritisation Mistakes to Avoid
Mistake #1: Team-Level Thinking
What appears high-impact for one team may deliver minimal value to the broader organisation. A customer service AI that reduces response time by 50% seems transformative to the support team, but if it only affects 2% of customer interactions, the organisation-wide impact is limited.
Always ask: "How does this create value for the entire business, not just this department?"
Mistake #2: Wishful Thinking About Implementation
Teams consistently underestimate implementation difficulty – especially cultural barriers. Technical feasibility is only part of the picture; you must also assess data integration complexity, system changes, cultural and process shifts, and any regulatory or compliance hurdles. This is why the reality check from Step 4 is so critical – your ease-of-implementation assessment must be grounded in facts, not optimism.
The Cultural Barrier Reality Check
Don't underestimate the cultural barriers that need to be overcome to make change happen. Technical solutions often fail due to human resistance rather than technical limitations.
Ask tough questions about each opportunity:
Will people actually use this, or will they find workarounds?
What behaviours must change, and how realistic is that change?
Who might resist this change, and do we have plans to address their concerns?
Are the required process changes compatible with existing workflows?
Cultural barriers often move opportunities from "easy to implement" to "complex change management required."
Making the Selection Decision
Select Two Opportunities Maximum
Choose one opportunity to pilot immediately and one to incubate for future consideration. This disciplined approach ensures adequate resources and attention for generating clear learnings.
Resist the Multiple Pilot Temptation
The instinct to pursue several pilots simultaneously almost always leads to under-resourced initiatives that fail to generate clear learnings. You end up with multiple inconclusive experiments rather than one definitive success or failure.
Start small and learn. One well-executed pilot teaches you more about AI implementation than three under-resourced experiments.
Documentation for Future Context
Record Your Reasoning
Document—or at least clearly remember—why you're not pursuing the other opportunities. This serves two critical purposes:
Maintains Focus: Clear reasoning helps resist scope creep and feature requests that pull attention from your chosen pilot
Provides Context: Future prioritisation decisions benefit from understanding previous logic, especially when conditions change
Sample Documentation Format:
For each opportunity not selected, capture:
Why it scored lower: Specific factors that reduced its priority
What would need to change: Conditions that might make it viable later
Timeline considerations: When it might be appropriate to revisit
Matrix Positioning Guide
Opportunities that offer high value and easy implementation are your best pilot candidates—they promise meaningful business impact without overwhelming complexity. Start here to demonstrate quick, visible results.
Those with high value but hard implementation should be incubated for later. Their potential payoff warrants investment once your organisation strengthens its technical and operational capabilities.
Meanwhile, low-value but easy-to-implement ideas can serve as small wins to build confidence, but don’t let them distract from high-impact work. Anything that’s both low value and hard to implement should be parked for now—save your resources for opportunities that truly move the needle.
Making It Practical
Scoring Approach: Use a simple 1-5 scale for each axis, but focus on relative positioning rather than precise numerical scores. The goal is clear differentiation between opportunities, not mathematical precision.
Decision Criteria: The highest combined score wins, but use your judgment about quadrant positioning. A 4,4 opportunity (high value, relatively easy) might be preferable to a 5,3 opportunity (very high value, moderate difficulty) depending on your organisation's current capability and risk tolerance.
Timeline Pressure: If you're under pressure to move quickly, bias toward ease of implementation. Success with a moderately valuable but achievable pilot builds credibility for tackling harder, higher-value opportunities later.
7. Building a Culture of AI Success
Successful AI delivery requires different approaches than traditional software development. Beyond just technical considerations, the right people, culture and operating rhythm are essential to ensure your AI initiatives thrive.
Building the Right Team: People Make or Break AI Projects
The success of your AI initiatives depends heavily on assembling the right team with both technical expertise and cultural alignment. Look for individuals who demonstrate curiosity, adaptability, and a willingness to challenge assumptions. The most successful AI teams blend diverse perspectives and bring complementary skills.
When selecting team members, prioritise those who demonstrate:
A growth mindset and comfort with ambiguity
Data literacy and critical thinking
Collaborative problem-solving skills
Business acumen alongside technical capability
Essential Roles for AI Delivery
Product Owner: Value and Outcome Accountability
The Product Owner maintains focus on customer value and business outcomes throughout the AI initiative lifecycle, acting as the bridge between technical capabilities and business needs. They must be empowered to make quick decisions and have strong stakeholder management skills.
Technical Lead: Architecture and Integration Reality
The Technical Lead ensures AI systems integrate effectively with existing infrastructure while meeting security, performance, and reliability requirements. They must balance innovation with practical implementation constraints.
Risk Lead: Safety and Compliance
The Risk Lead identifies and mitigates risks specific to AI applications, including bias, fairness, transparency, and regulatory compliance issues. This role requires both technical understanding and regulatory awareness.
Data Owner: Quality and Governance
The Data Owner ensures AI systems have access to appropriate, high-quality data while maintaining proper governance and compliance. They champion data quality as a foundation for AI success.
Finance Partner: Unit Economics and Sustainability
The Finance Partner ensures AI initiatives deliver sustainable financial returns and costs remain controlled throughout the lifecycle. They help develop business cases that properly account for AI's unique economics.
Delivery Manager: Coordination and Momentum
The Delivery Manager maintains project momentum and removes obstacles that could derail AI initiatives. They foster a culture of accountability and transparent communication.
Creating a Culture of AI Adoption
Technology alone doesn't drive transformation—culture does. Successful AI implementations require a supportive culture where:
Teams feel psychologically safe to experiment and learn from failures
Data-driven decision making is valued over intuition alone
Cross-functional collaboration breaks down traditional silos
Leadership visibly champions AI initiatives and their strategic importance
Operating Cadence for AI Success: Why Rhythm Matters
A consistent operating rhythm provides the structure needed to maintain momentum while allowing for the adaptation inherent in AI development. The right cadence creates accountability without stifling innovation, and ensures early identification of both problems and opportunities.
Weekly Delivery Stand-ups: Focus on immediate progress, blockers, and technical issues. These short, focused sessions maintain momentum and create regular touchpoints for rapid problem-solving.
Fortnightly Demo and Decision Meetings: Demonstrate working AI capabilities to stakeholders. These sessions build trust through transparency and provide natural decision points before further investment.
Monthly Business Reviews: Focus on business impact, cost management, and strategic alignment. These deeper reviews ensure AI initiatives remain connected to business outcomes and provide formal points for course correction.
Quarterly Business Model Impact Reviews: Evaluate how AI initiatives affect each of the nine building blocks. These strategic sessions ensure AI work contributes to fundamental business model innovation rather than just incremental improvements.
7. Building a Culture of AI Success
Successful AI delivery requires different approaches than traditional software development. Beyond just technical considerations, the right people, culture and operating rhythm are essential to ensure your AI initiatives thrive.
Building the Right Team: People Make or Break AI Projects
The success of your AI initiatives depends heavily on assembling the right team with both technical expertise and cultural alignment. Look for individuals who demonstrate curiosity, adaptability, and a willingness to challenge assumptions. The most successful AI teams blend diverse perspectives and bring complementary skills.
When selecting team members, prioritise those who demonstrate:
A growth mindset and comfort with ambiguity
Data literacy and critical thinking
Collaborative problem-solving skills
Business acumen alongside technical capability
Essential Roles for AI Delivery
Product Owner: Value and Outcome Accountability
The Product Owner maintains focus on customer value and business outcomes throughout the AI initiative lifecycle, acting as the bridge between technical capabilities and business needs. They must be empowered to make quick decisions and have strong stakeholder management skills.
Technical Lead: Architecture and Integration Reality
The Technical Lead ensures AI systems integrate effectively with existing infrastructure while meeting security, performance, and reliability requirements. They must balance innovation with practical implementation constraints.
Risk Lead: Safety and Compliance
The Risk Lead identifies and mitigates risks specific to AI applications, including bias, fairness, transparency, and regulatory compliance issues. This role requires both technical understanding and regulatory awareness.
Data Owner: Quality and Governance
The Data Owner ensures AI systems have access to appropriate, high-quality data while maintaining proper governance and compliance. They champion data quality as a foundation for AI success.
Finance Partner: Unit Economics and Sustainability
The Finance Partner ensures AI initiatives deliver sustainable financial returns and costs remain controlled throughout the lifecycle. They help develop business cases that properly account for AI's unique economics.
Delivery Manager: Coordination and Momentum
The Delivery Manager maintains project momentum and removes obstacles that could derail AI initiatives. They foster a culture of accountability and transparent communication.
Creating a Culture of AI Adoption
Technology alone doesn't drive transformation—culture does. Successful AI implementations require a supportive culture where:
Teams feel psychologically safe to experiment and learn from failures
Data-driven decision making is valued over intuition alone
Cross-functional collaboration breaks down traditional silos
Leadership visibly champions AI initiatives and their strategic importance
Operating Cadence for AI Success: Why Rhythm Matters
A consistent operating rhythm provides the structure needed to maintain momentum while allowing for the adaptation inherent in AI development. The right cadence creates accountability without stifling innovation, and ensures early identification of both problems and opportunities.
Weekly Delivery Stand-ups: Focus on immediate progress, blockers, and technical issues. These short, focused sessions maintain momentum and create regular touchpoints for rapid problem-solving.
Fortnightly Demo and Decision Meetings: Demonstrate working AI capabilities to stakeholders. These sessions build trust through transparency and provide natural decision points before further investment.
Monthly Business Reviews: Focus on business impact, cost management, and strategic alignment. These deeper reviews ensure AI initiatives remain connected to business outcomes and provide formal points for course correction.
Quarterly Business Model Impact Reviews: Evaluate how AI initiatives affect each of the nine building blocks. These strategic sessions ensure AI work contributes to fundamental business model innovation rather than just incremental improvements.
7. Building a Culture of AI Success
Successful AI delivery requires different approaches than traditional software development. Beyond just technical considerations, the right people, culture and operating rhythm are essential to ensure your AI initiatives thrive.
Building the Right Team: People Make or Break AI Projects
The success of your AI initiatives depends heavily on assembling the right team with both technical expertise and cultural alignment. Look for individuals who demonstrate curiosity, adaptability, and a willingness to challenge assumptions. The most successful AI teams blend diverse perspectives and bring complementary skills.
When selecting team members, prioritise those who demonstrate:
A growth mindset and comfort with ambiguity
Data literacy and critical thinking
Collaborative problem-solving skills
Business acumen alongside technical capability
Essential Roles for AI Delivery
Product Owner: Value and Outcome Accountability
The Product Owner maintains focus on customer value and business outcomes throughout the AI initiative lifecycle, acting as the bridge between technical capabilities and business needs. They must be empowered to make quick decisions and have strong stakeholder management skills.
Technical Lead: Architecture and Integration Reality
The Technical Lead ensures AI systems integrate effectively with existing infrastructure while meeting security, performance, and reliability requirements. They must balance innovation with practical implementation constraints.
Risk Lead: Safety and Compliance
The Risk Lead identifies and mitigates risks specific to AI applications, including bias, fairness, transparency, and regulatory compliance issues. This role requires both technical understanding and regulatory awareness.
Data Owner: Quality and Governance
The Data Owner ensures AI systems have access to appropriate, high-quality data while maintaining proper governance and compliance. They champion data quality as a foundation for AI success.
Finance Partner: Unit Economics and Sustainability
The Finance Partner ensures AI initiatives deliver sustainable financial returns and costs remain controlled throughout the lifecycle. They help develop business cases that properly account for AI's unique economics.
Delivery Manager: Coordination and Momentum
The Delivery Manager maintains project momentum and removes obstacles that could derail AI initiatives. They foster a culture of accountability and transparent communication.
Creating a Culture of AI Adoption
Technology alone doesn't drive transformation—culture does. Successful AI implementations require a supportive culture where:
Teams feel psychologically safe to experiment and learn from failures
Data-driven decision making is valued over intuition alone
Cross-functional collaboration breaks down traditional silos
Leadership visibly champions AI initiatives and their strategic importance
Operating Cadence for AI Success: Why Rhythm Matters
A consistent operating rhythm provides the structure needed to maintain momentum while allowing for the adaptation inherent in AI development. The right cadence creates accountability without stifling innovation, and ensures early identification of both problems and opportunities.
Weekly Delivery Stand-ups: Focus on immediate progress, blockers, and technical issues. These short, focused sessions maintain momentum and create regular touchpoints for rapid problem-solving.
Fortnightly Demo and Decision Meetings: Demonstrate working AI capabilities to stakeholders. These sessions build trust through transparency and provide natural decision points before further investment.
Monthly Business Reviews: Focus on business impact, cost management, and strategic alignment. These deeper reviews ensure AI initiatives remain connected to business outcomes and provide formal points for course correction.
Quarterly Business Model Impact Reviews: Evaluate how AI initiatives affect each of the nine building blocks. These strategic sessions ensure AI work contributes to fundamental business model innovation rather than just incremental improvements.
8. Conclusion
This playbook provides a comprehensive framework for implementing AI initiatives that create genuine business value rather than impressive technology demonstrations. The business model approach ensures every AI investment connects to customer outcomes and sustainable competitive advantage.
Key Principles
Start with Business Model, Not Technology: Always begin by understanding how your organisation creates, delivers, and captures value. AI is most powerful when it supports or changes fundamental business dynamics rather than just automating existing processes.
Maintain Discipline About Evidence: Ground every decision in observable data rather than assumptions. Be honest about what you know and what remains uncertain. Acknowledge limitations rather than overselling possibilities.
Focus on Outcomes, Not Features: Customers care about results, not technology. Measure success through customer value delivered and business objectives achieved rather than technical metrics alone.
Build Sustainable Capabilities: Invest in reusable AI infrastructure supporting multiple applications rather than building point solutions for every use case. This creates lasting competitive advantage while controlling costs.
Govern by Design: Build appropriate governance, ethics, and risk management into AI systems from the beginning. This protects both your organisation and customers while enabling confident scaling.
Implementing AI successfully requires patience, discipline, and continuous learning. Most organisations benefit from starting with the sprint to establish shared understanding and identify the most promising opportunities. From there, focus on executing one initiative exceptionally well before expanding to multiple simultaneous efforts.
Remember that AI implementation is as much about organisational change as technical development. Invest time in building the right team dynamics, governance processes, and measurement systems that support long-term success.
Your AI journey is unique to your organisation, customers, and competitive context. This playbook provides proven frameworks to guide your decisions, but success ultimately depends on your commitment to systematic implementation, honest evaluation, and continuous improvement.
The organisations that succeed with AI will be those that maintain focus on customer value while building sustainable capabilities for the long term. Start with clear business justification, maintain disciplined execution, and scale systematically based on evidence rather than enthusiasm.
Everyone else is asking "what can AI do?" instead, you're going to ask "how does my business actually work, and where can AI help?" The first question only gets you so far. Answering the second question gets you commercial advantage.
8. Conclusion
This playbook provides a comprehensive framework for implementing AI initiatives that create genuine business value rather than impressive technology demonstrations. The business model approach ensures every AI investment connects to customer outcomes and sustainable competitive advantage.
Key Principles
Start with Business Model, Not Technology: Always begin by understanding how your organisation creates, delivers, and captures value. AI is most powerful when it supports or changes fundamental business dynamics rather than just automating existing processes.
Maintain Discipline About Evidence: Ground every decision in observable data rather than assumptions. Be honest about what you know and what remains uncertain. Acknowledge limitations rather than overselling possibilities.
Focus on Outcomes, Not Features: Customers care about results, not technology. Measure success through customer value delivered and business objectives achieved rather than technical metrics alone.
Build Sustainable Capabilities: Invest in reusable AI infrastructure supporting multiple applications rather than building point solutions for every use case. This creates lasting competitive advantage while controlling costs.
Govern by Design: Build appropriate governance, ethics, and risk management into AI systems from the beginning. This protects both your organisation and customers while enabling confident scaling.
Implementing AI successfully requires patience, discipline, and continuous learning. Most organisations benefit from starting with the sprint to establish shared understanding and identify the most promising opportunities. From there, focus on executing one initiative exceptionally well before expanding to multiple simultaneous efforts.
Remember that AI implementation is as much about organisational change as technical development. Invest time in building the right team dynamics, governance processes, and measurement systems that support long-term success.
Your AI journey is unique to your organisation, customers, and competitive context. This playbook provides proven frameworks to guide your decisions, but success ultimately depends on your commitment to systematic implementation, honest evaluation, and continuous improvement.
The organisations that succeed with AI will be those that maintain focus on customer value while building sustainable capabilities for the long term. Start with clear business justification, maintain disciplined execution, and scale systematically based on evidence rather than enthusiasm.
Everyone else is asking "what can AI do?" instead, you're going to ask "how does my business actually work, and where can AI help?" The first question only gets you so far. Answering the second question gets you commercial advantage.
8. Conclusion
This playbook provides a comprehensive framework for implementing AI initiatives that create genuine business value rather than impressive technology demonstrations. The business model approach ensures every AI investment connects to customer outcomes and sustainable competitive advantage.
Key Principles
Start with Business Model, Not Technology: Always begin by understanding how your organisation creates, delivers, and captures value. AI is most powerful when it supports or changes fundamental business dynamics rather than just automating existing processes.
Maintain Discipline About Evidence: Ground every decision in observable data rather than assumptions. Be honest about what you know and what remains uncertain. Acknowledge limitations rather than overselling possibilities.
Focus on Outcomes, Not Features: Customers care about results, not technology. Measure success through customer value delivered and business objectives achieved rather than technical metrics alone.
Build Sustainable Capabilities: Invest in reusable AI infrastructure supporting multiple applications rather than building point solutions for every use case. This creates lasting competitive advantage while controlling costs.
Govern by Design: Build appropriate governance, ethics, and risk management into AI systems from the beginning. This protects both your organisation and customers while enabling confident scaling.
Implementing AI successfully requires patience, discipline, and continuous learning. Most organisations benefit from starting with the sprint to establish shared understanding and identify the most promising opportunities. From there, focus on executing one initiative exceptionally well before expanding to multiple simultaneous efforts.
Remember that AI implementation is as much about organisational change as technical development. Invest time in building the right team dynamics, governance processes, and measurement systems that support long-term success.
Your AI journey is unique to your organisation, customers, and competitive context. This playbook provides proven frameworks to guide your decisions, but success ultimately depends on your commitment to systematic implementation, honest evaluation, and continuous improvement.
The organisations that succeed with AI will be those that maintain focus on customer value while building sustainable capabilities for the long term. Start with clear business justification, maintain disciplined execution, and scale systematically based on evidence rather than enthusiasm.
Everyone else is asking "what can AI do?" instead, you're going to ask "how does my business actually work, and where can AI help?" The first question only gets you so far. Answering the second question gets you commercial advantage.
Continue Learning
Join Our Mailing List
Join Our Mailing List
Join Our Mailing List
The Innovation Experts
Our Offices
Get in Touch
© 2025 Bloch AI
The Innovation Experts
Our Offices
Get in Touch
© 2025 Bloch AI
The Innovation Experts
Our Offices
Get in Touch
© 2025 Bloch AI