Lens Two - Establishing the Center of Excellence
Author's note: Going Beyond AI is six lenses on building a transformation-ready organization, for CEOs and operators. If you missed Lens 1, read it here.
Nearly every company ran AI training this year. The budget cleared, the vendor showed up, and the room was fully engaged. People left energized—only to find that their newfound energy had nowhere to go.
Six weeks later, the only evidence the training happened is a line item on an invoice and a calendar invite nobody kept. The employee who attended still knows what they learned. The organization never did.
At Grow, every engagement starts in the same place: we find out where a company stores what it already knows. Most have nowhere, which is the whole problem this series is built around — transformation-ready organizations, one move at a time. The center of excellence is where that starts.
What an AI Center of Excellence Holds
A Center of Excellence (CoE) is a small, dedicated team that owns how a company adopts AI. It sets the standards, builds the shared tools, captures what works, and spreads it across the organization. It is not a bloated corporate department, and it is not an abstract budget line. It is simply a training ground with a floor under it—a shared repository where organizational knowledge accumulates.
A true operational CoE rests on four essential pillars:
The prompt library. What actually worked, written down, with the context that made it work.
Approved use cases. Where this is allowed, where it is not, and who decides the edge cases.
The failure log. What went wrong and why. This is the one nobody builds and the one that pays for the rest.
The judgment. What your best people know and have never had to write down. AI made writing it down worth doing for the first time.
Without these elements, you risk expensive operational blind spots.
For instance, in 2024 a Canadian tribunal held Air Canada responsible for a bereavement fare invented by its chatbot. The airline argued the AI was an independent legal entity. The tribunal swiftly rejected that "remarkable submission." The system had simply behaved exactly as it was built, because nobody had bothered to write down the boundaries of what it was allowed to say (Pinsent Masons, 2024).
Map Where Your Team Actually Is
Fluency moves along four dimensions, and your people sit in different places on each one. The Grow AI Mindset Framework scores all four:
Thought Partner. Whether AI is something you query or something you think with.
Tool to Workflow. Whether it is a thing you open or a thing wired into how work moves.
Productivity to Innovation. Whether it makes the current job faster or makes a different job possible.
Task to Outcome. Whether you hand it steps or hand it a result you want.
Run it on your leadership team before you run it anywhere else. A team that views AI as nothing more than a glorified search engine will only ever build search engines. No amount of software tooling can fix a flawed foundational assumption.
This is why a Center of Excellence functions as a management cadence rather than a static curriculum.
You are changing what people believe is possible, then giving those breakthroughs somewhere permanent to live.
IN OUR NEXT NEWSLETTER · LENS 3 · GOING BEYOND AI
Solving the Problem of the Laggards
The people falling behind are fearful or negative for reasons that make sense to them, and their direct manager is the only variable that reliably moves any of it.
Work with Grow:
We help CEOs and PE-backed companies redesign marketing for the AI era — leaner teams, sharper strategy, smarter execution, and growth systems built around measurable impact.
▸ Ready to scale smarter? Book a Growth Strategy Call → https://www.growpowered.com/contact-us
▸ Take our 3-Minute Marketing Performance Assessment today:
▸ Subscribe to our newsletter for more insights → https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7278079307564949505
Frequently Asked Questions:
What is an AI center of excellence?
A central place where an organization stores and improves its AI capability: the prompt library, approved use cases, a failure log, and documented judgment. It is a shared asset rather than a department or a budget line.
What should an AI center of excellence contain?
Four things. A prompt library of what actually worked and why, a list of approved and prohibited use cases, a failure log recording what went wrong, and the tacit judgment of senior people written down for the first time.
Why does an AI failure log matter more than a prompt library?
A prompt library makes a team faster. A failure log prevents repeat mistakes and creates the governance record that protects the company when an AI system produces something wrong.
Who should own an AI center of excellence?
One named person with a recurring weekly hour, not a committee. Committees are what organizations form when nobody wants to hold the accountability.
What are the four dimensions of AI fluency?
Thought Partner, Tool to Workflow, Productivity to Innovation, and Task to Outcome. Each describes a different way a person can relate to AI, and most people sit at different levels across the four.
What are the three stages of AI fluency?
Developing, Contributing, and Leading. Where someone starts matters less than whether they are progressing from one stage to the next.
Is AI training enough to build organizational capability?
No. Training is an event and the learning leaves with whoever attended. Capability requires somewhere for that learning to accumulate so the next person does not start from zero.
Comments