• AI SPRINT
  • Posts
  • [AI SPRINT] The Flywheel That Built Amazon (and Can Power Your AI)

[AI SPRINT] The Flywheel That Built Amazon (and Can Power Your AI)

This week: a simple way to choose AI projects that actually drive business results—straight from Amazon’s playbook.

There’s no shortage of places to apply AI right now.

Every leadership team I talk to is staring at a list of dozens of ideas for AI pilots, MVPs, or vendor pitches—and wondering: which ones will actually move the needle?

When I was at Amazon, we didn’t chase the next big thing. We used a system. And that system was built around one deceptively simple tool: the Flywheel.

Not the metaphor. Here’s the actual napkin sketch Bezos and his leadership team drew in 2001 to map out Amazon’s growth engine (just kidding, it’s AI).

At the center: their goal is “Growth”.
Around it: selection, customer experience, traffic, sellers, which all support each other.
And the top? A lower cost structure, which led to lower prices.

That top part—cost structure and low prices—was the lever.
And everything we launched had to show how it helped spin the flywheel faster.

What most people miss is this: the lever is driven by one thing: AI.

Amazon invested billions to make that their competitive edge. With generative AI and recent advancements, those same capabilities are now within reach of any company.

I use Amazon’s model to help companies prioritize AI investments, with a slight twist.

The AI Flywheel™ Method

If you want AI to be more than just a pilot program graveyard (or waste of time and money), start here:

1. Spin the flywheel
Choose projects that reinforce your core value loop—whatever drives your business forward. In manufacturing, that might be: Faster production → higher quality → more customer trust → increased orders.

2. Remove bottlenecks
Apply AI where human speed or cost is slowing down your system. These are your scale blockers.

3. Increase customer offerings
Use AI to enable services you couldn't deliver before: 24/7 support, personalized experiences, faster onboarding, or smarter recommendations.

Real Example: Launching Amazon Delivery in Tokyo

When I was leading Amazon’s Last Mile Innovation team, I launched delivery operations in Japan. We were forced to start “scrappy” due to outside pressures—little tech, just urgency. Drivers, packages, paper maps.

It broke almost immediately.

They were spending hours just figuring out how to plan delivery routes on paper. Some worked 16-hour shifts. It was chaos.

So we brought in AI. We trained it on Tokyo’s street grid, traffic patterns, and delivery data, then gave every driver an app with optimized real-time routing.

That one investment:

  • Removed the biggest bottleneck (manual routing)

  • Reinforced the flywheel (faster, cheaper delivery)

  • Created new offerings (same-day service at scale)

That’s how you pick a winning AI project.

Before you greenlight the next AI pilot, ask:

  • Does this project reduce cost or time across departments?

  • Will it remove a bottleneck that’s slowing down growth?

  • Does it help deliver something new or better to the customer?

  • Does it spin our flywheel—or is it just shiny?

If the answer isn’t clear, the value won’t be either.

Use the AI SPRINT Framework to Drive Fast AI Adoption

Here’s the model I use to make this all simple—it’s the project plan for your AI adoption.

  1. Spark Action: Educate and align leaders on AI.

  2. Position Company: Create your AI Flywheel to identify the best investments.

  3. Rally Employees: Get AI into daily use.

  4. Integrate: Put it into real, revenue generating workflows first.

  5. Enable: Build methods to nurture your employee adoption.

  6. Trailblaze: Launch new AI-powered services to your customers.

If you apply it, your AI efforts won’t just reduce costs.
They’ll create leverage.

In the future, there won’t be AI-powered businesses and traditional businesses—only AI-powered businesses and obsolete ones

The Takeaway?

Amazon didn’t scale through big bets.
It scaled through clear levers and repeatable systems.

It scaled with AI.

If your org is serious about AI, start by asking:

What’s your flywheel—and where’s your lever?

Then build the systems to push, again and again.

About Trent: Trent Gillespie is an AI Keynote Speaker, CEO of Stellis AI, former Amazon leader, and advisor on building AI-Native, AI-Enabled businesses. Book Trent to speak to your group or book a call to discuss using AI within your business.

Tell us what you thought of today's email.

Login or Subscribe to participate in polls.

Did someone forward this newsletter to you? If you're not already signed up, you can subscribe to AI SPRINT™ for free here.