What if the operating system for AI transformation already exists?

Many of the leadership team I talk to is asking the same question right now. What does an organization look like when AI is part of how we build, decide and work? Which roles change? How fast can we move without breaking trust?

And since agile died a few years ago, at least according to LinkedIn, the assumption is that we need something completely new. A new framework. A new operating model. Maybe even a new certification.

But what if we don’t? When I started building my AI Transformation Playbook, I went looking for the foundation an AI transformation needs: learning fast under uncertainty, keeping people at the center, adapting as the technology moves. I kept landing on the four values of the Agile Manifesto. They are not new, but AI is truly putting them to the test.

In the playbook, I’ve translated them into four trade-offs in plain language. Then I asked what each one means on two tracks: the technology track, where we learn what AI can actually do for us, and the people track, where we learn whether people can see the strategy, trust the process and find their part in it.

People and conversations over tools and licenses

(From the manifesto: individuals and interactions over processes and tools)

Most AI programs today start with the tools. That leaves an opening for those who start with people instead. The organizations that win with AI will not be the ones with the best models. They will be the ones whose people learn fastest.

On the technology track, tools and models are replaceable. The one you choose today will be outdated within a year. So design around the workflow and the person using it, not around the tool.

On the people track, trust, relationships and participation decide adoption. Most AI rollouts invest heavily in licenses and lightly in conversations. It should be the other way around.

Real use over business cases

(From the manifesto: working software over comprehensive documentation)

On the technology track, real use in a limited setting beats a business case. A demo should show behavior and evidence, not slides about potential.

On the people track, people believe what they experience. One hour with a real task beats ten communication plans. If you want people to trust AI, let them try it on their own work, safely, before you tell them what it will do for them.

Co-creation over sign-off

(From the manifesto: customer collaboration over contract negotiation)

On the technology track, domain experts and users co-design and evaluate every experiment. They know where the edge cases live. The AI team usually doesn’t.

On the people track, it gets interesting. Every employee has an unwritten contract with their organization: I do good work, you give me meaningful work, fair treatment and a future. AI touches every line of it. That contract has to be renegotiated openly, and the people affected need a voice before decisions are made, not after.

Learning over plan precision

(From the manifesto: responding to change over following a plan)

On the technology track, a Minimum Viable Plan and short decision cycles beat a roadmap. Treat model choice as a changeable decision. The landscape will move faster than any plan you write.

On the people track, transitions are paced to the people, not the project. Name what is ending before you sell what is beginning. And let the plan adapt to readiness, not the other way around.

The part most people skips

In Scrum, the five values are the part most people skim on their way to the events and artifacts. In the playbook, they have become the behaviors we expect from leaders and teams. With AI, they might be the most important part:

• Commitment to outcomes and to the people affected, not to a solution
• Focus on a few meaningful problems, connected to the strategy
• Openness about results, errors, costs and what AI means for jobs
• Respect for expertise that is changing and for people who move at a different pace
• Courage to stop experiments, to report failures and to make honest promises

Courage is the one I would underline twice. It takes courage to stop a pilot the executive team loves. It takes courage to report that the model hallucinated in front of a customer. And it takes courage to tell people honestly what AI will mean for their jobs, including the parts you don’t know yet.

So, does it already exist?

I think it does. Not as a framework you can buy, but as a set of values you have to live. Agile was never about sprints and stand-ups. It was about learning faster than the world around you, with people at the center. AI raises the pace, and that makes the values more relevant, not less.

The new part is not the operating system. It is running it on two tracks at once, technology and people, at a speed we have never seen before.

For something that died a few years ago, agile looks surprisingly ready for the future.

This is the first part of my AI Transformation Playbook. More will follow. If you’d like a preview, reach out.

Which of these four trade-offs do you think AI will challenge the most in the years ahead?

Image credit: unsplash.com