EP 259 What Is an AI Flywheel? (And Why your Pilots Stall)

Most leaders talk about AI in terms of pilots, projects, and one-off tools. In this solo episode, host Susan Diaz explains why that mindset stalls adoption - and introduces the idea of an AI flywheel: a simple, compounding loop of audit → training → personalized tools → ROI that quietly turns experiments into momentum across your whole organisation.

Episode summary

Susan opens by contrasting how most organizations approach AI - pilots, isolated chatbots, a few licences to “see what happens” - with how enduring companies build flywheels that compound over time.

Borrowing from Jim Collins’ Good to Great, and examples like Amazon’s recommendation engine, she reframes AI from “one big launch” to a heavy wheel that’s hard to move at first, but almost impossible to stop once it’s spinning.

She then introduces her AI flywheel for organizations, built on four moving pillars:

1. Audit - reality-check where AI already lives in tools, workflows, risks, and guardrails.
   
   

2. Training - raise the floor of AI literacy so more people can safely experiment.
   
   

3. Personalised tools and workflows - move beyond generic prompts into department- and workflow-specific systems.
   
   

4. ROI tracking - measure time saved, errors reduced, risk reduced, and adoption so the story keeps getting funded.
   
   

Instead of a linear checklist, these components form a loop - each turn of the wheel making the next easier, and creating an unfair advantage for organizations that start early.

Finally, Susan adds the outer ring: human-first culture and governance as the operating system around the flywheel - psychological safety, champions and mentors, and values like equity that ensure AI momentum doesn’t quietly recreate hustle culture or leave people behind.

She closes with practical questions any leadership team can use this week to start their own AI flywheel.

Key takeaways

Projects start and end. Flywheels don’t. Treating AI as a string of pilots and vendor launches creates start–stop energy. Designing a flywheel turns every experiment into input for the next win.

A flywheel is heavy at first - but gains unstoppable momentum. Like a giant metal train wheel, it needs a lot of initial force, but each full turn adds speed. AI works the same: early experiments feel slow, compounding learning later feels unfairly fast.

The AI flywheel has four core pillars:

• Audit - map current tools, workflows, risks, and guardrails; discover hidden wins and power users.

• Training - treat AI like financial literacy: a minimum viable level for everyone so they can ask better questions and prompt more effectively.
  
  

• Personalised tools & workflows - stop asking “Which LLM?” and start asking “Which steps in this 37-step process should AI do?” Workflow first, tool second.
  
  

• ROI tracking - measure time saved, errors reduced, faster time to market, risk reduction, and % of AI-augmented workflows so leaders keep investing.

Culture is the operating system around the flywheel. Without psychological safety, people hide experiments. Without support, power users burn out. Values like equity matter: who’s getting trained, who has access, and who you’re helping reskill. Governance should feel like guidance, not punishment.

You don’t build an AI flywheel in a day. You start with one audit, one workflow, one dashboard that makes things more transparent - and commit to one small centimetre of momentum at a time.

Episode highlights

[00:02] Why “we’re piloting a chatbot” is not a strategy.

[01:34] Flywheel 101: the train-wheel analogy and why momentum beats one-off effort.

[03:19] Amazon’s recommendation engine as a classic business flywheel.

[05:02] Applying Jim Collins’ Good to Great flywheel lens to AI initiatives.

[05:30] From big bang ERP-style AI projects to small, compounding loops.

[08:00] Introducing the four pillars: audit, training, personalised tools, ROI.

[08:53] Audit as reality check: surfacing hidden wins and DIY power users.

[11:14] Training as “raising the floor” of AI literacy.

[14:08] Workflow-first thinking and the myth of the single all-powerful agent.

[17:33] ROI stories: error reduction, faster time to market, and risk reduction.

[20:19] Culture as outer ring: psychological safety, champions, values in action.

[23:06] Starting your flywheel: three questions for your leadership team.

Use this episode as a design tool, not just a definition.

Grab a whiteboard with your leadership team and map:

1. Where are we already auditing, training, personalising tools, and measuring ROI - however informally?

2. Where is the wheel broken, or missing entirely?

3. What’s one centimetre of movement we can create this quarter - one audit, one workflow, one dashboard - to start our AI flywheel turning?

Connect with Susan Diaz on LinkedIn to get a conversation started.

 

Agile teams move fast. Grab our 10 AI Deep Research Prompts to see how proven frameworks can unlock clarity in hours, not months. Find the prompt pack here.