The Most Dangerous Agent Is the One Failing Quietly
The AI failure that hurts isn't the dramatic one. It's the agent that's been quietly, confidently wrong for three weeks in a workflow nobody's watching. Here's where to look.
Give Every AI Agent a Name
The fastest way to lose control of your AI isn't a rogue model — it's a dozen anonymous agents nobody owns. Naming them is where governance actually starts.
The Three Questions That Are Your Whole AI Governance
The governance-gap piece struck a nerve, and the follow-up question was always the same: okay, so what do I actually do? Here's the answer — and it's smaller than the enterprise crowd wants you to think.
Please Stop Putting AI in Everything (From a Pro-AI Operator)
I'm a giant believer in AI for operations. That's precisely why forced, bolted-on AI drives me up the wall. If a task takes thirty seconds and works, making it twenty-nine isn't efficiency — it's theater.
AI didn't shrink operations — it made operations the surface where growth actually happens.
The takes said AI would gut operations first. The data says the opposite. Ops isn't the cut line — it's the surface where everything you build actually scales.
Nobody Decides to Build a Replacement. They Default Into One.
Nobody sits down and decides to replace a person with AI. They default into it — the moment they build for a job they've never done and a user they never talked to. And AI agents punish that mistake harder than anything before them.
The COO Is Quietly Becoming the Company's AI Owner. Good.
Ask who owns AI and most people say the CTO. It's landing somewhere else: the COO. Because the hard part of AI was never the model — it's everything operations already owns.
Fractional Was the On-Ramp. Full-Time Is the Destination.
Why the operator role keeps getting rented — and why the smartest founders eventually buy.
The fractional operations wave was one of the healthiest things to happen to growing companies in years. It let founders who couldn't yet justify a full executive hire get real operational leadership on a schedule they could afford. It worked. I've done it, and I've watched it change how companies run.
But I want to say something that cuts against my own business model: for a lot of companies, fractional was never the destination. It was the on-ramp — the thing that finally showed them the role was load-bearing.
You Don't Have an AI Problem. You Have a Governance Gap.
Everyone deployed AI agents. Most companies can't govern them. Governance isn't a model setting — it's decision rights, escalation paths, and accountability. In other words, operations.
I Watched AI Come for My Job — So I Learned to Wield It
A field report from the messy, exhilarating middle of reinventing an operations career around AI.
There's a specific kind of quiet that settles over you when you realize the tools being built across your industry could do your job. Not help you do it better. Do it instead of you.
I felt that quiet a while back. I work in tech, in operations — the connective tissue that keeps a company actually running. And I watched the market lean toward tools that weren't really designed to make operational people faster or sharper, but to make them optional. When you spend your days close to how these tools get built, you start to notice the assumption baked into a lot of them: that the work of operations is mostly automatable. Sitting with that trajectory is a strange thing.
I could have spent that season being anxious about it, and some days I was. But somewhere in there I made a different decision, and it's the decision this whole article is really about: I decided I would rather out-learn the shift than wait to see where it left me.
AI Isn’t Coming for the Cheapest Jobs First. It’s Coming for the Most Structured Ones.
The first wave didn’t go the way the headlines predicted
If AI simply displaced the lowest-cost roles first, we’d expect the economics to be obvious. Instead, an MIT analysis put the share of roles where automation is actually economically viable at roughly 23% — meaning humans remain the cheaper option in more than three-quarters of them. The capability exists in far more places than the cost case does.
The reversals tell the same story. Klarna cut roughly 700 customer-service roles on the expectation that AI would absorb the work, then moved to rehire when service quality slipped and customers noticed. Amazon’s “Just Walk Out” shopping technology, widely described as AI doing the work, leaned heavily on remote human reviewers behind the scenes. And Forrester now projects that about half of AI-attributed layoffs will be quietly reversed — often refilled at lower cost rather than admitted as a miss.
When Does AI Become Quality Management vs. Micromanagement?
There's a line between measuring work and monitoring people. Most teams cross it by accident.
A pattern I’ve seen play out more than once: a team rolls out a helpful internal tool. Nothing dramatic — a simple way to track tasks inside a platform they already use. A manager sends a quick request, the recipient taps back a status — done, blocked, in progress. Clean, fast, genuinely useful. Everyone likes it.
Then a later update adds one small field. When you mark a task complete, a new box appears: how long did this take you? No explanation, no note about why. It just shows up one morning, quietly asking everyone to put a number on their minutes. And the whole feel of the tool changes.
It's Not the Data. It's the Question.
The analyst's job didn't shrink. It moved.
It’s tempting to think that tools which answer instantly reduce the need for human judgment. The opposite is true. When answering becomes cheap, the value shifts entirely upstream — to framing the question, defining the terms, and deciding what would actually count as evidence. The model handles the retrieval. The human owns the inquiry.
So when someone tells me the AI “gave the wrong answer,” my first question is rarely about the model. It’s: what exactly did we ask it? More often than not, the data was fine. The question was the thing that needed the work — and that, not the answer, is where real analytical expertise lives.
The Most Valuable Thing AI Can Do for Your Operation Isn’t Replace People. It’s Remember Them.
A friend sent me Pam Didner’s recent piece in Inc., “Forget Prompting. The Most Important AI Skill of the Next 5 Years Is Something Else.” Her argument stopped me because it lands exactly where I’ve spent my career living: the gap between what your best people know and what your organization can actually reproduce.
Her thesis is simple. The next real AI skill isn’t better prompting or vibe coding. It’s writing a skill file — a one-page document that tells an AI tool how to do a specific task your way, every time. Not a style guide. Not a prompt library. One task, your standards, your format, written down so the work runs the same whether or not the person who mastered it is in the room.
I’d put it more bluntly, from an operator’s chair: a skill file is institutional knowledge that finally stops walking out the door.
AI-Augmented Operations: A Field Guide for Founders Scaling Past Early Stage
There's a moment in every growing company where the founder realizes the operating system they've been running on — the spreadsheets, the founder-in-the-loop approvals, the tribal knowledge living in three people's heads — has quietly stopped scaling.
The team is bigger. The compliance surface area is bigger. The customer base is bigger. And the founder is spending more time on operational maintenance than on the strategic moves that actually grow the business.
This is the moment when most founders start thinking about hiring an operator. It's also the moment when a lot of them get distracted by a different question: “Can AI just do this instead?”
The honest answer is more interesting than either yes or no.
AI cannot replace operations. But AI can — and should — fundamentally reshape what operations looks like inside a company scaling past early stage. The companies that figure this out get a multiplier on every dollar they spend on operations leadership. The companies that don't usually end up paying for the lesson twice.
This guide is for founders trying to figure out the right way to integrate AI into their operations — without making the mistakes the early movers just made at scale.
What CEOs Get Wrong About "AI-Augmented" Teams
Every week I hear some version of the same conversation.
A CEO has decided the company is going “AI-first.” They’ve purchased a suite of tools. They’ve announced it in an all-hands. They’re expecting efficiency gains within the quarter.
Six months later, they’re frustrated. Adoption is patchy. Half the team is using the tools, half isn’t. The efficiency gains haven’t materialized. And somewhere in a Slack channel nobody checks, there are three different AI workflows doing three versions of the same thing, none of them talking to each other.
This is not an AI problem. It’s an operational design problem. And it stems from a handful of mistakes I see CEOs make almost every time.
What I Actually Do in My First 90 Days
Most executives spend their first 90 days listening. I spend mine fixing.
That’s not a brag. It’s a structural reality. When a founder brings in a fractional COO or Chief of Staff, they’re not hiring someone to observe. They’re hiring someone because something — often several things — has quietly broken, and the company is scaling faster than its operating system can handle.
The “listen first” advice exists for a reason: premature action based on incomplete information is a real failure mode. But there’s another failure mode nobody talks about. It’s the executive who spends 90 days in discovery while the team keeps bleeding, the process debt keeps compounding, and the founder keeps wondering when they’re going to see a return on the hire.
Here’s what I actually do.
Good Process Is Invisible
Say the word "process" in the wrong room and watch what happens.
Screens don't go down. Eyes don't light up. What you get instead is a specific kind of stillness — the kind that means everyone in the room just added something to their mental list of things they're going to have to tolerate.
That reaction is data. Not about process. About what people believe process is.
They believe it's a rulebook. A stack of requirements. A manager's way of adding friction to work that was flowing just fine. They think of process as the ten-page document that nobody reads, the approval chain that slows everything down, the form that has to be filled out before anything can move.
Why Your AI Rollout Is Really a Trust Audit
There’s a moment in almost every AI rollout that nobody talks about.
It’s not the launch meeting. It’s not the training session. It’s the moment — usually in the first few weeks — when a leader gets a suggestion from a team member, pauses, and says: “let me just check what the AI says about that.”
What happens next determines everything.
If the AI confirms the employee’s suggestion, the leader nods and moves on. If the AI says something different, the leader goes with the AI.
The employee notices. They always notice.
And what they’ve just learned isn’t anything about the AI tool. What they’ve learned is that their judgment — their experience, their contextual knowledge, their read of the situation — is now subject to machine review. That the AI is the authority. That they are not.
That’s not an AI problem. That’s a trust problem. And it didn’t start with the rollout.
AI in Service of Culture: The Communication Problem Most Leaders Get Wrong
The first piece made the operational case for structure inside a flat organization. The second piece argued that structure only works when leadership has the trust to let go. The third piece showed the ROI — that healthy culture, built on clear structure and real trust, outperforms the S&P 500 by roughly 2,000% in cumulative returns over time.
This last piece is about the question every leader I work with eventually asks: where does AI fit into all of this?
The short answer is that AI belongs in service of culture, not in place of it. It’s a supporting role, not a replacement. And the reason most companies get the AI rollout wrong has very little to do with the technology — and almost everything to do with how leadership communicates it.