Align → Activate → Amplify → Accelerate → Govern

As the checklist header spoiled already, OpenAI’s guide says the winning pattern is simple and repeatable: Align → Activate → Amplify → Accelerate → Govern.

Consider this a “minimum viable operating system” for an AI-first company.

And there are some great case studies for each phase of this model:

Align

Make AI adoption non-negotiable and visible.

As OpenAI says: “Employees adopt change faster when they clearly see how new AI initiatives enhance their skills, enable more meaningful work, and contribute to their company’s competitive advantage. Leaders play a critical role in driving this alignment by explicitly communicating the purpose behind AI initiatives, demonstrating their commitment, and actively supporting employees throughout the transition.”

We ​discussed before​ how Moderna’s CEO sets the expectation that employees use ChatGPT ~20 times a day.

OpenAI’s CFO Sarah Friar shares her own usage, normalizing experimentation, and making the finance (!) team one of the most advanced AI adopters at OpenAI.

That’s what role-modeling looks like.

Even just simply setting a calendar (or ChatGPT!) reminder to share something you’ve discovered to get your team or company experimenting is key.

Activate

Fix the training gap.

The guide quotes ​McKinsey research​ which shows that nearly half of employees lack support to use generative AI, while ranking training as the #1 success factor.

This echoes the findings in the recent viral MIT report, which highlighted that ​95% of AI pilots fail​, in large part because teams haven’t been properly trained for it.

And there are plenty of positive examples. The San Antonio Spurs ​lifted AI fluency​ from 14% to 85% by embedding training into daily work.

Guided by HR Tech chief ​Marlene de Koning​, PwC Netherlands ​built a network of AI champions​ early on in their ​AI implementation​, another best practice from the paper. (​Members​, her ​Masterclass on Copilot adoption​ is a must-watch.)

And experimentation, another success driver, was the name of the game in J&J’s “​Let a Thousand Flowers Bloom​” approach.

(A great starting point for Activating AI is our ​Executive AI Boot Camp​. We also can adapt this for your goals and platforms – ​let’s discuss​.)

Amplify

Stop reinventing the wheel.

The whitepaper is quite blunt: “The fastest way to scale AI impact is to stop solving the same problems in silos.”

Build a central hub, share wins weekly, and keep an active champions network.

This can be very simple. ​Lead with AI PRO​ member ​Carlo Benigni​ saves his prompts in a Notion database for easy copy-pasting.

PwC created a ​shared library​ of successful AI use cases and prompts to help employees learn from each other and improve their AI skills.

And at Asana, there’s a ​common Slack channel​ to discuss what they’re using AI for, sharing wins, but importantly also failures.

Accelerate

Reduce time-to-production.

As I ​shared before​, it’s critical to let employees drive AI adoption, as they know best where in their roles AI can make an impact.

For that, the whitepaper says, it’s important to have utmost clarity on which tools to use, what’s allowed, and where approvals are needed.

A dedicated function for this makes sense. Estée Lauder’s GPT Lab ​collected​ 1,000+ ideas and prototyped high-value GPTs, and then helped scale the highest-value use cases.

And BBVA ​created​ an “AI network” to fast-track proof-of-concepts into production while ensuring smooth collaboration across departments.

For acceleration to happen in the right direction, keep in mind ​AI change management​ best practices like prioritizing high-ROI functions, aligning interests, and ensuring ethical use of AI.

Govern

“Moving fast doesn’t mean ignoring risks.”

This is not a “Wild West” of AI experiments, but a clearly guardrailed context where we’re moving fast, but not breaking things.

Interestingly, the paper is light on information here. It’s recommending playbooks with “safe-to-try” rules, quarterly reviews, and lightweight approvals, but lacks case studies or deeper insights.Perhaps this is the area where even AI leaders still stumble, but fortunately, we do have ideas here:

WSJ tech writer ​Alexandra Samuel​ says to ​invest in secure AI tools​ that ensure data privacy and security, reducing concerns about data misuse.

And ​Rebecca Hinds​ from Glean says “Level 5” AI companies ​focus early​ on developing guidelines and principles.

The Bottom Line: Leading AI in 2025

With this whitepaper, OpenAI published what’s expected of you today.

Align → Activate → Amplify → Accelerate → Govern is a “minimum viable operating system” for modern leaders.

Your 30-day catch-up plan:

• Set the bar in public. Commit to two new AI-supported workflows following the ​GED-RT model​ per team this month and visible daily usage metrics. Share your own examples weekly.
• Stand up the “middle.” Name a small ​AI council​ and spin up a single AI hub (policies, prompts, playbooks, wins) so teams stop reinventing the wheel.
• Train where work happens. Run role-specific clinics for managers where you can teach them prompting, delegation, and judgment using your real documents and data. (​We can help​.)
• Make momentum visible. Showcase wins on a cadence, track time-to-production, and reinvest budget and time into teams delivering measurable impact.

(Feels overwhelming to do by yourself? Enroll for our October 3rd or November 14th ​Executive Boot Camp​, where you learn and implement AI in your work, team, and organization.)

Until next week,

Daan