AI Adoption Masterclass: How to AI-Pill Every Single Employee
You might have seen Ramp’s announcement that 99% of their employees are using AI and that every single employee gets their own AI employee.
Or you might’ve seen Uber’s CTO explaining how engineers nearly burned through the entire annual AI usage budget in just 3 months, and that he’s now going back to the drawing board.
And there you are, sitting and wondering:
How do I make my employees become AI-pilled, automate their workflows, work 10x faster, and focus on new initiatives with the free time AI gives them?
How do you make your employees obsessed with AI and stay competitive with the companies already ahead?
How do you motivate them the right way?
How do you move them from years of ingrained traditional workflows to genuinely embracing new ones?
Because there’s no competing with a small AI-native company that has a swarm of agents working 24/7 and producing outputs at a quality and speed you simply can’t match with a traditional team.
In the next 10 minutes, I’m going to give you all the answers: based on conversations I’ve had with leaders across industries, and what I’m practically doing inside my own company right now:
The budget objection is a distraction.
I’ve heard it in boardrooms, in leadership offsites, in 1:1s with managers who should know better... “we’d love to do more with AI but the cost...”
A Claude Pro plan is $20/month. A GitHub Copilot seat is $19. An agent that automates a full weekly workflow costs less per month than one team lunch.
And apart from your engineering team - if you know what you’re doing - your costs shouldn’t be crazy high!
Claude Cowork, Dispatch, Routines, etc are so awesome in getting tons of work getting done without costing anything extra.
If your company isn’t AI-native right now, money is not the reason.
Your competitors aren’t just hiring faster or spending more on ads. They’re building companies where every employee (not just the technical ones) is running their own agents, automating their own work, and compressing what used to take a week into a Tuesday afternoon.
What’s Your Problem Is You Haven’t Burnt Your Boats Yet
Sending an email that says “we’re going to become AI-native and every employee should be using AI”... good luck seeing any real change three months later.
You’ve to burn the boats. Issue that memorandum. Make your managers accountable. Give them a proper AI transformation plan with concrete, measurable outcomes:
“By Q3, your agents should be handling 80% of the work employees are currently doing across every department. That’s the target. Put it in writing.”
“Before anyone asks for new headcount, they show you proof that an AI agent genuinely cannot do that job. That’s the new bar.”
“And the overarching mandate: by Q3, every single person in this org is managing their own AI systems and agents. Not still experimenting and exploring.”
Because your role ( and your team’s role) is going to evolve into one thing: managing an army of agents.
The people who figure that out early will be the ones your company is built around. The ones who don’t will be replaced by the agents themselves.
And before you start making this change, I want to save you from the mistakes that probably you’re making right now:
The Mistakes Most Companies Are Making
Mistake #1: Talking the talk but not walking it
If you’re a founder or CEO and you’re not personally AI-maxxing, you will not build a culture that does. You can’t give a town hall about AI transformation and then go back to writing emails the way you wrote them in 2021.
Look at Tobi Lütke at Shopify. Look at Dharmesh Shah. Look at Gary Tan.
These are some of the most successful entrepreneurs in the world. They don’t need to work another day in their lives... yet they’re building with AI like their survival depends on it.
They’re not the people who forwarded an article to their leadership team and called it a strategy. They’re in the tools. They’re building things.
They’re posting the actual outputs. And their orgs reflect that - you can feel it in how their teams talk about AI, in how fast they ship, in the quality of what they put out.
When the person at the top is visibly operating in a different productivity tier, the entire org gets permission to follow. More than permission, they feel the pressure to keep up. If you’re not doing that yet, start there. Before the all-hands. Before the taskforce. Get your hands dirty, figure out what works, and then you’ll have something real to say.
Mistake 2: Trying to Do Everything at Once
You feel the FOMO. You see Ramp, you see Shopify, you read the LinkedIn posts and something in your stomach says we’re behind. So you push hard, set aggressive timelines, and try to make the entire transformation happen in one quarter.
What you end up with is a castle built on sand.
When you force the pace, your team doesn’t actually transform... they “perform transformation.” They’ll build you something demo-able because that’s what the pressure demands. The demo looks great. The all-hands goes well. And then three months later you’re standing at point zero wondering why nothing actually changed, why the workflows are still manual, why the agents aren’t running, why the team quietly went back to the way things were.
Desperation is a terrible input for any strategic decision and AI transformation is no different. The companies getting this right aren’t the ones moving fastest. They’re the ones moving most deliberately. Phases with clear outcomes. Small wins that build internal confidence. A foundation solid enough that when you scale on top of it, it holds.
Calm down. Plan it properly. The urgency is real but panic is not a strategy.
Mistake 3: Believing “Agents Don’t Work” or “AI Can’t Do My Job”
You’ll hear this constantly. You might even believe it yourself after a few failed attempts. But this narrative is almost entirely false... and understanding why it’s false is the difference between companies that figure this out and companies that don’t.
When an agent fails, it is never the AI’s problem. It is a context problem. A skill problem. A “you” problem.
Here’s what’s actually happening:
Humans are extraordinarily bad at explaining how they do things because most of our expertise lives below the surface. It’s instinct.
Pattern recognition built over years that we’ve never had to articulate because we’ve never had to teach a machine to replicate it.
You know how to write a great outreach email. But if someone asked you to write down every decision you make while writing it the ones about tone, timing, word choice, what to never say... you’d struggle.
Because you’ve never had to think about it that way.
Alex Hormozi once broke down what it takes to write a great viral hook. It’s a 22-step process. Twenty-two. Every amateur thinks the instruction is “write me a viral hook.” Every expert knows there are two dozen decisions that happen before the hook is any good.
Your agents are no different. Telling an agent “do this for me” and expecting great output is like handing a new hire a one-line job description and expecting them to operate like a ten-year veteran on day one.
What you actually need to build are “process delegation pathways.”
Take the task. Do it yourself, slowly, out loud.
Break it down to its smallest possible components.
Document your instincts (and concrete proof): what makes a good output, what makes a bad one, what you’d never do and why.
Build the guardrails. Define the edge cases. Give the agent examples of 10/10 work and 3/10 work and explain the difference.
Then test it. Score it. Find where it breaks. Refine the instructions. Run it again.
This is not an AI problem. This is a process documentation problem that AI has made impossible to ignore. The teams that crack it don’t have better models than anyone else. They just taught their agents better than everyone else did.
Mistake #4: Celebrating token consumption instead of outcomes
This is the mistake you’ll encounter when your team is already using like a pro...
I’ve watched something spread through certain tech companies that I think is genuinely counterproductive: leaderboards for AI usage volume. Shoutouts for whoever sent the most prompts this month. The Meta-style race to see who burns the most compute.
This is the wrong thing to cheer for.
Token maximization is the AI equivalent of celebrating hours worked instead of problems solved. It creates a culture of performance... people running AI tools visibly rather than running them well.
What you actually want to track is outcome per initiative.
Can someone point to a specific piece of work, say what they did with AI, and say what it produced? Not “I used Claude for 4 hours.”
More like: “I automated the proposal process. It used to take 3 days per client. It now takes 40 minutes. We’ve sent 3x more proposals this quarter.”
That’s the unit of value.
The best practitioners are actually token-minimizers. Tight prompts, precise instructions, fast outputs. If you’re going to build a leaderboard, build one that tracks value created per initiative — not volume consumed.
Here’s How to Actually Transform Your Company
Step 1: Understand your archetype
Not every company AI-transforms the same way. A professional services firm has different leverage points than a SaaS company. An ops-heavy business has different starting points than a media company.
Before you pick a stack or start building agents, understand what kind of company you are: where time is spent, where decisions happen, where output is most constrained. The companies that skip this step jump straight to building agents before they understand their own workflows well enough to automate them. The agent breaks. They blame the AI & Adoption stalls.
Step 2: Transform in phases
Phase 1: Workflow automation. Take the repeatable, manual, rule-based work your team does every week and automate it (across every department). Reports, summaries, data pulls, draft generation, inbox triage. This phase doesn’t require AI sophistication. It requires someone willing to sit with a team member for two hours and map every step of what they actually do. Start there. The wins compound fast and they build the internal confidence that makes the next phase possible.
Phase 2: Build a second version of yourself. Every person in your org has a body of knowledge... how they handle a certain type of customer, how they write a certain communication, how they approach a specific problem. Phase 2 is about externalizing that into an agent that can do a version of that work without them doing it every time. A CS lead builds an agent that handles tier-1 inquiries the way she would. A growth manager builds an agent that runs his weekly SEO analysis the way he’d run it. And the list goes on.
Phase 3: Full monotonous agentic systems at scale. This is where a company looks genuinely different from the outside. Whole categories of work happening continuously, without a human kicking it off each time... leads being enriched as they come in, content being drafted and queued, reports being generated and distributed. The human role shifts to quality control, refinement, and exception handling. Time goes entirely toward things that actually need a human.
The Most Important AI Hire You Can Make Right Now
If your org doesn’t have the internal capability to do all the things above, the instinct is to call a consultancy. Don’t. You’ll get a beautiful slide deck in four months, a hefty invoice, and zero internal capability to show for it. The transformation lives in the deck. Not in your team.
What you actually need is a full-time AI Transformation Lead. Someone who ships in days, not quarters. Someone who gets their hands dirty alongside your teams, builds the systems, and leaves internal capability behind as they go. Not a theorist. A builder who happens to understand organisations.
But here’s what I think works even bette and what I’m personally leaning towards.
Across your departments ( marketing, design, engineering, operations) hire part-time contractors from the creator ecosystem on X or wherever you can find them.
Find the people who are already publicly building, automating, and shipping in those specific domains. Give them a full month sprint. Embed them with their field peers inside your team.
The reason this works so well is simple: an expert in workflow automation is not the same as an expert in marketing workflow automation. The outputs are completely different.
A rockstar growth operator who lives and breathes AI-powered marketing knows the tools, the edge cases, the what-actually-works in that domain in a way a generalist automation consultant never will.
When you put that person in a room with your marketing team, the learning transfer is immediate. They’re speaking the same language from day one.
So the structure I’d recommend: one AI Transformation Lead (actual builder) sitting across the organisation, holding the vision, managing the process, connecting the dots. And underneath that, domain-specific contractors running focused sprints inside each department — the AI-native marketing expert with your marketing team, the AI-native engineer with your engineering team, and so on.
There are many ways to structure this. But this is the one I’d back right now.
How to Incentivise the Transformation (This Is What Most Leaders Skip)
Getting employees to genuinely go AI-native isn’t about mandates. It’s about making it the most rewarding thing they can do.
Build the leaderboard... but track the right things
Create a visible, company-wide leaderboard. Not for token usage. For transformation impact.
Track things like: number of workflows automated, hours saved per week, initiatives launched using AI, before/after comparisons on output volume. Make it public. Update it regularly. Let people see where they stand.
Nothing moves people faster than visibility. When someone sees their name climbing a leaderboard ( or notices they’re near the bottom) that’s a more powerful motivator than any all-hands slide about the importance of AI.
Publicly recognise the builders
When an employee builds a workflow that saves 5 hours a week, call it out. In the all-hands. In Slack. In your internal newsletter. Give it a name — “the (Employee Name) playbook.” Let them demo it to the team. And add it to your context brain.
People are wired to want their work seen and admired. Nothing gives someone more of a bump than being recognized by leadership for building something that actually matters. Make AI-building the thing that earns that recognition, and you’ll have a line of people who want to be next.
Incentivise workflow creation directly
Go further than recognition. Tie it to real rewards. Because there’s nothing that motivate people than incentives.
If someone automates a workflow that demonstrably saves the team X hours per week, reward them. Bonus. Time off. A meaningful shoutout that goes on record. Some companies are even letting employees who build high-impact internal agents take a small percentage of the value they generate. The specifics matter less than the signal: we take this seriously, and we put something on the table to prove it.
The goal is to make building AI workflows feel like one of the highest-leverage career moves someone can make at your company. Because it is. And once your best people start seeing it that way, the culture shifts on its own.
Create an internal AI transformation track
Track each employee’s “AI-pilled” journey explicitly and make managers accountable:
Where are they today... still doing everything manually?
Running their first automations?
Operating with a full personal agent stack?
Make progression visible. Give it a name.
Create internal levels if you want: AI Starter, AI Builder, AI Native. Let people self-report and get their managers to validate. Run a weekly check-in on where the org is moving as a whole.
This isn’t bureaucracy. It’s a forcing function. When there’s a visible map and people can see where they are on it, they move toward the next level. Especially when moving to the next level means recognition, rewards, and the respect of the people they work with.
What the Other Side Looks Like
The companies that come out of this period in a strong position won’t be the ones who ran the best AI pilot. They’ll be the ones where using AI became so embedded in how work gets done that not using it became the weird thing.
Where every person (from the CEO to the newest coordinator) is running their own systems, building their own agents, and shipping work that would’ve taken a team to do two years ago.
That shift doesn’t happen with a policy. It happens when leadership walks it, when the vision is loud and specific, when the people who build get recognised for it, and when there’s a way to track that the org is actually moving.
The competitors who figure this out will not announce it. You’ll just notice, one day, that they’re doing 10x the output with the same headcount. By then, catching up is a different conversation entirely.