Context Is the Moat

It’s eleven at night and a second-year associate at a twelve-person strategy boutique is staring at a problem that, two years ago, would have cost her a frantic call to a partner. The client wants a read on whether a mid-market acquisition makes sense by morning. She doesn’t call anyone. She opens a model, and in ninety minutes she has a competitive teardown, a clean diligence framework, three valuation approaches with their assumptions laid out, and a list of the questions a good firm would ask. It is genuinely good work — the kind that used to justify a senior’s day rate and a junior’s all-nighter.

Here is what should make every founder of such a firm sit up: the boutique across town has the exact same associate available to it, at the exact same price, tonight.

For a century, professional services sold one thing above all others — judgment that was scarce. You hired the sharpest people, trained them into partners over a decade, and sold their thinking by the hour. The scarcity was the business. That scarcity is now collapsing, not because people got worse, but because a very large slice of what we called judgment just became a utility.

So the urgent question isn’t how do we adopt AI. It’s the one underneath: when intelligence is a utility, what’s actually still ours?

In the AI era, intelligence commoditizes and context doesn’t — so the moat moves from who has the smartest people to whose context compounds, and who owns it.

This essay is the case for taking that sentence literally, and for treating it as the most important strategic fact of the next decade in professional services.

1. The old moat dissolves

Talent was always a leaky moat; we just didn’t say so out loud. Your best people walked out the door every night and, eventually, out of the firm. The judgment you spent years building lived in their heads, and when they left, it left with them. The entire apprenticeship model — pyramids of juniors absorbing partner judgment over years — is an elaborate, expensive workaround for the fact that your core asset had legs.

AI doesn’t patch that leak. It does something stranger, and to see it clearly you have to name precisely what the technology is. The model is an encyclopedic senior associate. It has read every framework your profession ever published, absorbed every best practice, internalized every methodology, and it will recall the correct one, correctly applied, in seconds — at three in the morning, on a holiday, without ego, politics, or burnout. Ask it to structure a market entry analysis, run first-pass diligence, draft the methodology section, frame the issue the way a top-tier firm would, and it does that layer extraordinarily well.

It is, in other words, the best-read associate your firm will never have to train. And it bills the same to you and to everyone you compete with.

So be exact about what just got commoditized. Not judgment wholesale — the encyclopedic layer of it: the book-knowledge, the frameworks, the methodological scaffolding that quietly justified a great many senior hours. That layer is now abundant. And abundance, in any market, is priced at zero margin. This isn’t speculative. The early-2026 substitution wave is already visible — firms re-routing work to cheaper models with no loss in quality, and at least one legal-AI firm publicly reporting that a fine-tuned open model beat a top frontier model on its own benchmark at a fraction of the cost. Capable reasoning is getting cheaper and more uniform, not scarcer.

The talent moat isn’t eroding gracefully — its encyclopedic half is being arbitraged in real time.

But push that associate to the edge of its competence and you hit a wall — and the wall is the whole story. It has read about ten thousand turnarounds. It has lived through none. It has every framework and no scar tissue. It will never have gut instinct, street smarts, the school of hard knocks, or the plain wisdom that only comes from years of being right, being wrong, and learning to tell the two apart in advance — not because the technology can’t improve, but because this kind of judgment is earned in the world, not read off a page. It knows what the textbook prescribes when a target’s numbers don’t quite reconcile. It does not have the prickle on the back of a twenty-year partner’s neck that says something here is off before she can articulate why — the instinct, built from a hundred deals, that makes her ask the one question that kills the deal and saves the client.

It has read ten thousand turnarounds and lived through none. Every framework, no scar tissue.

That earned, experiential judgment is the part of your people the model cannot touch. And here is the trap it sets: today, that wisdom lives almost entirely in your people’s heads. Which means it still walks out the door every night — and, when a partner retires, out of the firm forever, uncaptured and unrepeatable.

2. What AI can’t commoditize is you

So two things sit permanently outside the model’s reach: your firm’s specific context, and your people’s earned wisdom. They look like different problems. They’re actually the same one, and seeing that is the hinge of the entire thesis.

Start with context. The model doesn’t have the four hundred engagements your firm has run. It doesn’t know that an approach exactly like the one you’re about to recommend quietly failed for a client exactly like this one in 2021 — and why. It doesn’t know the methodology your founders spent a decade refining, the kind of work your firm has learned to decline, the way your best people actually pressure-test a number. It doesn’t have your deliverables, your decision history, the tacit “how we really do it here” written in the margins of every deck. The model is general by design — that generality is its genius, and precisely why it can never be your advantage.

Now the wisdom — and watch how it connects. The partner’s instinct that “something is off,” the hard rule that a certain client profile always overruns, the reason your firm walks away from a particular engagement shape: that isn’t magic. It’s experiential judgment that simply hasn’t been written down yet. It’s tacit context.

Your partners’ instinct isn’t magic. It’s context that hasn’t been written down yet.

Which points straight at the move most firms are missing. The same encyclopedic associate that commoditized your frameworks is the best instrument you have ever owned for capturing the rest. It can sit with a principal and turn “we always sanity-check the working-capital assumption” into a documented standard. It can interview the partner who’s three years from retirement and encode four decades of pattern-recognition before it walks out the door for good. It can watch how decisions actually get made across your engagements and surface the unwritten playbook nobody ever had time to write. AI hands you the encyclopedic layer for free — and, used deliberately, it is how you finally convert the wisdom that used to evaporate into context that stays.

That is the inversion at the heart of the AI era. For a century, the firm was the people and the work product was the output. Now the output is increasingly generated, and the firm is the context — the accumulated corpus of judgment, plus the captured wisdom, that turns a generic model into your model. Same engine, everywhere. Different fuel. The fuel is the firm.

3. Context is the only asset that compounds

Here’s why this reframing matters more than it first appears: context is the rare firm asset that compounds — and compounding, given enough time, beats everything.

Talent depreciates and walks; the wisdom that used to retire with your partners now compounds in an asset you keep. Brand is slow and diffuse. Methodology can be copied off your website. Context does the one thing none of them do: it accrues, and it accrues faster the more you use it. Two flywheels turn here, and they reinforce each other.

The first is the quality flywheel. Every engagement you run, every decision you log, every deliverable you produce is more context. More context makes the next answer sharper, more grounded, more unmistakably yours — which wins better work, which generates more context. The firm that captures its own output gets compoundingly better at its own job. The firm that lets it wash away re-derives the same insight, from scratch, forever.

The second is the switching-cost flywheel. The more of your firm’s context lives in a system that genuinely works, the more that system is worth to you and the harder it becomes to give up — not through lock-in, but because the accumulated value is real and irreplaceable. A firm sitting on ten years of compounded, structured context holds something a new entrant cannot buy and a competitor cannot copy. It can only be grown, and only by the firm that lived the engagements.

Picture two firms that open their doors on the same day, with equally bright people. Firm A treats its work as something to deliver and forget. Firm B treats every engagement as a deposit into an asset. For a year, you can’t tell them apart. By year three, Firm B answers faster and lands harder because it stands on everything it has ever done. By year seven, the gap isn’t a gap — it’s a category difference, and Firm A can’t close it by hiring, because the thing it’s missing isn’t talent.

There is no lateral hire for ten years of compounded context.

That is the literal definition of a moat: an asset that deepens the longer you operate and that no amount of a rival’s capital lets them skip ahead on. Talent never gave you that. Context does.

4. The sort has already started

None of this is a forecast. It’s underway, it’s mechanical, and it’s quietly sorting the entire industry into two piles.

Walk into any two firms of comparable size and you’ll find the same raw material — years of engagements, decisions, hard-won lessons — in two completely different states.

In the first, it’s exhaust. The diligence model from the deal that taught everyone a lesson is in someone’s downloads folder. The reasoning behind the firm’s signature methodology is in the founder’s head. The “we always check X” is an oral tradition that survives only as long as the person who remembers it. Every project starts near zero because nobody can find — or trust — the prior version. Every departure is a withdrawal from an account the firm never funded.

In the second, it’s capital. The same material is captured, structured, and feeding back into the work. The retiring partner’s judgment is encoded before she leaves. New people get good fast because they’re standing on the firm’s accumulated context, not re-learning it. Every engagement is a deposit.

Start those two firms at parity and let compounding run. The distance between them doesn’t widen in a straight line — it widens the way compounding always does: slowly, then suddenly. AI doesn’t cause this sort. It violently accelerates it, because the firm with captured context now has a general-purpose engine that can finally exploit all of it at once.

The sort isn’t smart firms versus dumb ones. It’s firms that treat their judgment as exhaust versus firms that treat it as capital.

The uncomfortable implication: plenty of brilliant firms are bleeding their context into the void right now and calling it busy.

Being smart was never the same as being defensible.

5. A moat you rent is not a moat

There’s a catch, and it’s the one most firms will get exactly backwards. The advantage only holds if you own the context. And ownership is precisely what today’s path of least resistance quietly gives away.

Here’s the trap, dressed up as convenience. You adopt a tool, and your firm’s knowledge starts accumulating inside it — your engagements, your decisions, your captured wisdom — on the terms you happened to accept. Plenty of vendors do offer real ownership: no training on your data, clean export, governance you control. But it is never the default; it’s the thing you have to require, in writing. Where you don’t, the context still compounds — it just compounds inside someone else’s asset, on terms that can change. The day the pricing moves or the terms shift, a moat you don’t own can be repriced or revoked, and you have little recourse, because you were a tenant the whole time.

Renting intelligence is fine. Renting your context hands your supplier the deed to your moat.

So the real question splits in two, and a firm has to answer both: Is our context captured? and Do we own it — portably, governed, ours to move and ours to keep? Sovereignty over your context isn’t an IT preference or a security checkbox. It’s the line between building an asset and furnishing one to a landlord.

This is the problem Aether exists to solve — making a firm’s own context an owned, compounding asset rather than exhaust it leaks or equity it rents. But the principle stands no matter what you build it with:

Own the context, or don’t call it a moat.

6. The repricing

Follow the logic to its end and you land somewhere that should change how these firms are valued.

For a century, a professional-services firm has been worth, roughly, its people plus its pipeline — a multiple applied nervously, because everyone in the room knew the assets took the elevator down every evening and could resign by email. That discount was rational. The value was mobile and rentable.

As context becomes the moat, the basis of value should travel with it. The durable question shifts from how many senior people do you employ to how much owned, compounding context have you built — and how defensibly do you hold it. A firm whose advantage walks out the door at night is worth one thing. A firm whose advantage compounds inside an asset it owns is worth another — and only the second kind is built to survive a world where raw intelligence is a commodity anyone can buy.

I’ll mark that as thesis, not fact: the repricing hasn’t fully arrived, and anyone quoting you precise multiples for it is guessing. But the direction isn’t seriously in doubt, because it’s just two forces we already understand — compounding and ownership — pointed at an industry that has never had to think in either term.

Soon, acquirers won’t ask who you employ. They’ll ask what you own — and whether it keeps growing without you.