Last week I wrote about why mandating AI utilization is a losing game – the gap between executive vision and day–to–day execution, and how most organizations are trying to close it with top–down pressure instead of bottom–up enablement.
A few of you wrote back and said some version of: “okay, but what does it actually look like when someone fills that gap?”
So I went and found one.
From ChatGPT Curious to AI Enablement Lead
My friend Doran works at Cloudflare. His title today is AI Enablement and Operations Lead – a role that didn’t exist before he created it a few months ago.
He got there by doing something that sounds straightforward until you watch most people fail to do it: he started using AI on his own, got really good at it, and then made himself available to everyone around him who wasn’t.
Doran started with ChatGPT the day it launched, eventually moved to Claude and Gemini, and just kept going. But here’s what separates him from the thousands of other early adopters who played with these tools and went back to their day jobs: he understood something most people miss entirely.
Sure, he could have set up a lunch and learn or sent around a “10 Ways to Use ChatGPT at Work” doc. Instead, Doran chose a different path entirely and positioned himself as the translation layer.
“Engineering Is Speaking French to Non–Technical People”
That’s how Doran describes the core problem. Comprehension is the bottleneck – most people inside large organizations hear “AI” and end up somewhere between “magic wand” and “threat to my job” and neither of those framings does anything useful for adoption.
What Doran figured out is that the job is spending less time talking in platitudes about what’s possible and more about digging into real business problems that are AI-solvable and then showing people exactly how to do it. His opening line was always some version of “tell me what’s taking you too long” – and then he went and helped people solve it.
Show Me Your Work, I’ll Show You a Better Way
Here’s what his role actually looks like day to day:
- Training sessions on internal AI tools built around live problem–solving – someone brings a real workflow, Doran works through it with them in real time. They leave with a working solution, not a slide deck.
- Building lightweight internal agents – a webpage draft agent that takes a brief and produces a first pass. A video subtitle translation agent that eliminated hours of manual work. These are specific, annoying problems that people were already complaining about – the kind of small wins that build credibility fast.
- Solving use cases in the room – instead of building something for a team and handing it over, he sits with them, builds it in front of them, and walks them through why each decision was made. Next time, they do it themselves.
This is the part most companies get wrong. They hire a Head of AI or buy an enterprise platform and expect adoption to follow. Adoption comes from one person sitting next to another person and saying “here, let me show you how this solves your specific problem” – and that part requires a human, every time.
What This Means for Your Organization
If you read last week’s issue and recognized your company in the mandate trap, here’s the practical takeaway: what most organizations need more than another AI strategy document is someone like Doran – a person who has already built real muscle memory with these tools and is willing to use it in service of everyone else’s problems.
That person tends to look something like this:
- They use AI daily and have the receipts to show for it – the kind of fluency that comes from months of iteration, failure, and refinement on real work. Weekend tinkering produces a different kind of person entirely.
- They can translate between technical possibility and business need – and critically, they lead with a colleague’s language before they reach for the tool’s vocabulary. The technology earns its place in the conversation, rather than dominating it from the start.
- They lead with curiosity about other people’s problems – they show up with questions about your workflow before they show you what a tool can do, and they earn trust before they earn scope.
- They start small and make it visible – a single workflow fixed, a single hour of manual work eliminated, a single team that now ships something faster than everyone else. That’s how credibility compounds inside an organization.
If the Mandate Is Already on the Table
Some of you aren’t in a position to decide whether or not to make the call – you already made it. And now you’re sitting with a utilization dashboard that’s underwhelming, a culture that has started to associate the word “AI” with surveillance or pressure, and a leadership team asking why the investment hasn’t moved the needle yet.
This is a recoverable situation, but the move forward is probably the opposite of what instinct suggests. Doubling down on the mandate – more tracking, more required trainings, more all-hands messaging – tends to deepen the resistance rather than dissolve it. Remember, the problem was never awareness, it was always trust and relevance and those two things respond to a completely different kind of intervention.
A few things that can actually work at this stage:
- Pull back the scope and go deep somewhere specific. One team, one workflow, one genuinely painful problem. Put your Doran on it for a few weeks. A visible win in one pocket of the organization travels faster than a company-wide initiative.
- Change what you measure. If you’re tracking logins and prompts submitted, people feel counted rather than supported. Shift toward outcomes – time saved, steps eliminated, a process that used to take three days and now takes one.
- Give people a way to tell you what’s actually in the way. Low adoption is almost always a signal that something specific is broken. A short, candid conversation with a handful of disengaged people will surface more insight than any adoption dashboard.
The mandate created urgency. What closes the gap now is proximity – someone with credibility and patience sitting with people inside their actual work, the same way Doran does.
The Role That Didn’t Exist
Doran didn’t wait for Cloudflare to post a job listing. He made himself so useful in this capacity that the role formed around him. That’s not an accident – that’s what happens when someone fills a gap that everyone feels but nobody has gotten around to putting words to.
If you’re a leader reading this: look for this person. They’re probably already in your organization, and you might not know it yet. They’re the one whose team keeps shipping faster than everyone else. They’re the one other people go to when they’re stuck, even when that’s not formally their job. They might not have “AI” anywhere in their title – yet.
If you’re the person who’s been doing this informally: call it what it is. Pitch it. Write down the three workflows you’ve already fixed, the hours you’ve already saved, the problems you’ve already translated into solutions, and bring that to your manager as a proposal. What Doran did at Cloudflare is a pattern that’s emerging everywhere right now, and the people who make it legible inside their organizations are the ones who get to shape what comes next.
Your Takeaway This Week
This week: spend fifteen minutes doing one of these three things.
If you’re a leader who hasn’t made the call yet – make a short list of who on your team has been building AI muscle. Who has the credibility, the curiosity, and the cross–functional relationships to be a Doran for your organization? Go have a conversation with that person. Ask them what they’ve been working on. You might already have the bridge you need.
If you’re working through a mandate that’s stalled – pick one team, one workflow, one problem. Take the pressure off the broader initiative for a moment and put everything behind making one thing work visibly and well. Start with the boring, aggravating stuff and let the success of that story do the traveling for you.
If you’re the builder doing this work informally – write down three specific things you’ve helped someone else do faster or better using AI in the last six months. That’s the beginning of a pitch. Give the informal work a title, and give yourself permission to go make it official.
The bridge was always going to be a person (or a tiny team of people) – someone who understood both the work and the people doing it. The good news is that person is probably closer than you think.
If you found this issue useful, share this issue with a friend. I wrote this newsletter for us – the people who are building the future at work.
– Annie
Annie Tsai 
Comment On This Post: