The Bet Against Accountants Is Going to Lose
There is a bet being placed on the accounting profession in 2026. It is going to lose.
The PE consolidators are buying firms on a thesis of cost reduction. The offshoring providers are pitching cheaper labor as the path to margins. The AI vendors are pitching headcount cuts as the value proposition. Three different players. One identical bet. Shrink the labor line, expand the margin.
They are betting against the accountant.
That bet is wrong, and the next few years are going to prove it wrong.
I have been writing about this shift for a long time now. Seema Amble and Steph Zhang at Andreessen Horowitz published two pieces last week that name what I have been seeing from inside the profession. Seema’s is called “Is Software Losing Its Head?” Steph’s is called “From System of Record to System of Intelligence.” Read together they describe how the moats in enterprise software are moving. Out of the database. Into the intelligence layer above it. Their conclusion lines up with what those of us building inside accounting have been saying for years.
That migration is the single most important thing happening in our industry. It is also the thing the consolidation wave is pricing in backwards.
The Moats Are Moving
The classic moats of enterprise software are getting hollowed out by agents.
UI does not matter when the agent does the clicking. Muscle memory does not matter when the agent does the workflow. Frequency of use does not matter when the agent shows up daily and the human does not. Salesforce spent twenty years building a fortress on those moats. The fortress is now a commodity.
The new moats live somewhere else. Proprietary data the product uniquely causes to exist. Operational logic. Context. Execution in the real world.
Look at that list and tell me it does not describe accounting.
There is one more thing the agentic world demands, and it is the thing most people miss. Agents need exactly what a new hire needs. Explicit rules. Permissions. Process definitions. Context for the decisions they are being asked to make. Without those, an agent cannot act safely, and a new hire cannot either. That requirement makes the people who carry the rules and the context more valuable, not less. The accountant who knows why this client closes the books the way they do, and what the auditor flagged three years ago, and how the CFO wants the variance commentary phrased, is the source of the logic the agent needs. Strip that out and the agent is dangerous. Keep it in and the accountant gets leverage they have never had before.
That is the moat shift in one sentence. The value used to live in the software. It is moving into the people who know how the work actually gets done.
Accounting Is Real World Execution
The signed audit opinion is execution in the real world. The filed tax return is execution in the real world. The reconciled close that gets sent to the board on Monday morning is execution in the real world.
Software vendors do not produce those artifacts. Accountants do.
Software vendors have been trying to climb into our industry for two decades. None of them have produced an audit opinion. None of them have signed a tax return. None of them have stood in front of a CFO and explained why a number that looked clean six weeks ago needs to be restated. That is the work. That is what we do. That is the moat.
The reason this matters in the agentic era is that pure software has nothing to put its name on. A SaaS company ships a feature. An accountant ships a deliverable that carries legal weight, fiduciary weight, and personal weight. You can sue an accountant. You cannot sue a chatbot. The accountability lives with the human, and so does the trust, and so does the relationship, and so does the price.
That is not a problem AI removes. It is a problem AI makes more valuable to solve.
The Intelligence Layer Lives in the Accountant
Steph’s piece extends the argument. The system of record, the database, becomes infrastructure. The system of intelligence, the layer above it, becomes the product.
Translate it into our world. The ERP becomes plumbing. NetSuite, Sage Intacct, QuickBooks, all of them. The intelligence layer is what sits above the ERP. It orchestrates the GL, the bank feeds, AR, AP, tax, payroll, time tracking, audit workpapers, and the client conversation. It reads, reasons, and acts on the data. That is where the value migrates.
That layer does not exist inside the ERP. It lives in the accountant’s head. It always has.
This is what we filed a patent on earlier this year with the Client Intelligence System. The premise is straightforward. Every engagement at Wiss should compound. Every exception we resolve should become context that informs the next engagement. Institutional knowledge should not live trapped in someone’s head or buried in a software system you cannot extract it from. It should be captured at a level that is impossible to pull out of a generic database.
This is the layer the profession has needed for fifty years and never had the technology to build.
Institutional Memory as Product
Steph wrote a sentence that should be tattooed on every managing partner’s wall.
Institutional memory becomes something a company can actually ship.
Worth sitting on that for a minute. The thing our profession has bled out for fifty years is the thing that becomes the product.
Every senior who retires takes a decade of client context with them. Every staff who quits takes the workflow they built in their head. The pipeline collapse made it worse. We are losing accountants faster than we are replacing them, and we are losing institutional memory continuously. The turnover has been brutal for as long as I have been in this profession, and the cost of it has always been invisible because nobody could measure what walked out the door.
That is the part the agentic era changes.
When the intelligence layer captures the logic, the context, the playbook, and the judgment patterns of a senior accountant, the firm finally owns the asset it has been losing every two years to attrition. The successor inherits the institutional memory instead of starting over. The client gets continuity instead of a relearning curve. The firm gets a moat that compounds instead of resets.
That is what shippable institutional memory means in our world.
The Pie Is Getting Larger
Now look at the consolidation thesis again.
The PE rollups assume labor is the cost line to shrink. The offshoring shops assume labor is interchangeable. The AI vendors assume software replaces labor. All three assume the value lives in the database and the workflow.
It does not. The value lives in the accountant.
Steph also published the data point that breaks the cost cutting argument. Companies adopting AI in their sales and marketing functions are spending more on labor, not less. ROI per dollar is rising. The pie is getting larger. Software is doing more and the team is doing more, simultaneously.
That is the contrarian case our industry has not made out loud yet. AI does not shrink the labor pie in accounting. It expands it. Because the leverage on a great accountant goes up, not down.
The harvesting path priced in a world where labor was the problem. The building path is showing up in a world where labor is the product.
Pick a Side
We are at the sharpest inflection point in a hundred years of accounting.
Most of the writing on the industry comes from inside it, which makes it incremental, defensive, and at times self serving. The Andreessen pieces matter because they describe where the value is migrating from a seat that watches dozens of industries at once. They are not telling us anything practitioners building in this space did not already know. They are validating it with a framework the market will recognize.
The value is moving up the stack. Out of the tools. Into the people. The accountants who learn to own the intelligence layer become more valuable, not less. The ones who do not become the commoditized input in someone else’s intelligence layer.
That is the bet at Wiss Labs. We deploy AI to amplify accountants, not to replace them. The accountant stays in control. The AI handles what does not require judgment and surfaces the parts that do. The senior signs every deliverable. The institutional knowledge gets captured instead of walking out the door.
The harvesting path is going to lose. The building path is going to win.
Pick a side.

Couldn't agree more with you on this, Paul.
Many software developers have failed attempting to automate a skill that we all know cannot be automated. Will tasks become easier to execute? Sure. That'll just mean more time to deliver more meaningful work for more clients.
Paul Peterson is exactly right - the bet against accountants is going to lose. I’d extend the same argument to wealth managers and advisors broadly.
AI is not eliminating the value of trusted professionals. It’s increasing the leverage of the best ones.
Anyone will be able to generate outputs with AI.
Very few will be able to interpret nuance, coordinate complexity, navigate human behavior, and help clients make decisions with confidence. That layer becomes more valuable in an AI world, not less.
The firms using AI to compound institutional intelligence around their people are building.
That’s the side I’d bet on too.