Fully staffed product teams: 10 years ago vs. now

Ten years ago, I worked with lots of "fully staffed" product teams: 4-6 engineers, dedicated QA, scrum masters, product owners, AND product managers. Teams of 8-10+ people coordinating constantly.

Now? I believe we can be much more ambitious on similar budgets by changing the default team structure entirely.

My recommendation for companies who can budget to "fully staff": small teams built around a core of 3 engineers, all managing AI agents.

Here's my ideal engineer composition:

• 1 senior+ tech lead orchestrating multiple agents
• 1 senior/mid-level partner engineer, also managing agents
• 1 junior in an apprentice role, learning both software engineering discipline AND agent management with heavy guidance

This isn't about replacing humans with AI. It's about redirecting their effort to make more progress.

3 engineers with the right agents and workflows can deliver more value than yesterday's team of 6.

And then you can make another similarly effective team with the other 3!

Why this works:

AI is an amplifier of both good and bad - so good tools, workflows, and clear measurable goals have the most direct impact.

But there's a non-obvious advantage that's critically important: smaller teams have much lower coordination costs. It's easier to maintain alignment and shared understanding of a focus area with ~5 people vs. ~10.

Critical success factors:

1. Product engineering mindset - Train and incentivize engineers to think like product people, not just coders. They need exposure to customers to make the right decisions.
2. Testing as a core discipline - When agents can generate code faster than humans can review, a testing mindset becomes non-negotiable. It's worth it to have a dedicated tester skilled in both automated AND exploratory testing. Quality will be a key differentiator now that writing the code itself is trending towards "free".
3. Strong supporting cast - You still need all of Marty Cagan's product "trio" engaged fully on each product team: engineering, product management, and design. But with stronger product engineers, these other roles can support multiple small teams of engineers without spreading too thin. Especially when teams focus on related areas.

Big picture:

In 20 years building software products, I've never worked for a company with enough development capacity to match its ambition. But this model can get you closer.

Managing multiple teams brings its own challenges. Team Topologies is my go-to framework here. You'll need platform teams as the number of product teams multiply.

And yes, you'll invest meaningfully in AI tooling.

But the return? You can finally get closer to matching your ambition with your capacity.

The question for you: How are you organizing teams now that managing AI agents effectively is as crucial as managing people?