{"slug": "company-does-not-have-processes-it-has-habits", "title": "Company does not have processes. It has habits", "summary": "A new essay argues that AI agent startups are targeting the wrong scarce resource: the real bottleneck is not labor or model capability but the undocumented, habit-based nature of business processes. The author contends that the most valuable output of an AI automation project is the written specification of the process, which alone enables faster onboarding and error detection, and that much of the work does not require a model.", "body_md": "There are dozens of well-funded companies building the same AI agent for the same back office, and they are not wrong about the capability. They are aimed at the wrong scarce resource. The constraint was never labour, and it was never the model.\n\nStart with the honest part, because it is the part most people writing about this skip.\n\nAutomating business processes with language models is one of the most crowded categories in software right now. AI teammate, AI employee, digital worker, agentic operations, autonomous back office. Dozens of well-capitalised companies, several of them very good, are building something that looks a lot like what we build. The demo is always the same shape: watch it read the email, find the record, work out what is being asked, make the change, write the note.\n\nAnd the demo works. That is the thing people who are sceptical of this category keep getting wrong. It genuinely works now, in a way it categorically did not two years ago, and it is going to keep working better on a schedule nobody in this market controls.\n\nWhich is exactly why nobody is going to win this on capability. Everyone is building on the same handful of frontier models. The gap between the best agent framework and the median one is measured in months and closing. If your entire thesis is that your agent is smarter than their agent, you have built a business on a rented advantage that depreciates every time somebody else ships a checkpoint.\n\nThe differences that will actually decide this category are not in the agent at all. They are in two things that turn out not to exist inside the companies buying it.\n\nHere is what happens every single time you go to automate a business process.\n\nYou ask for the documentation. You get a wiki page, written two or three years ago, by someone who has since left, describing a system that has been replaced once. It covers the happy path. The happy path is maybe sixty percent of volume.\n\nSo you go and sit with the person who actually does the work. And what you find is not a process. It\nis an oral tradition. There are perhaps forty exceptions, none written down, most of which resolve to\na sentence like *\"well, for that customer we do it differently, because of the thing that happened\nwith the freight claim.\"* The rule is real, it is load-bearing, it is correct, and it exists in\nexactly one skull. When that person is on holiday, the exception is handled wrong and nobody notices\nfor a quarter.\n\nThis is not a sign of a badly-run company. It is what every company is. Processes accrete out of individual judgements made under time pressure and then get copied by whoever sits nearest. A business does not have processes. It has habits, plus a document that describes an idealised version of some of them.\n\nSo the automation project stops being an automation project. Before you can encode the decision, you have to find out what the decision actually is, which means finding the person who holds it and getting them to say it in a sentence that survives contact with a schema. Half of every engagement is archaeology.\n\nWhich leads somewhere uncomfortable to say in a sales meeting.\n\nThe most valuable artefact produced by an AI automation project is the written specification\nof the process. It would have been worth commissioning even if you never deployed the model.\n\nWith a real specification you can onboard in days instead of months. You can find the two rules that contradict each other, which is usually the actual cause of the error rate somebody blamed on carelessness. You can compare two branches doing the same job three different ways. You can outsource it, price it, or decide to stop doing it. None of that requires a model. All of it was previously impossible because the process was distributed across people's memories.\n\nAnd then the second uncomfortable thing arrives immediately behind it. Once the process is actually written down, a large fraction of it turns out not to need a model at all. It is lookups, joins, threshold comparisons, and a state machine. It is code. Boring, deterministic, testable, cheap code that will run identically in five years.\n\nThe model earns its place on the ambiguous slice: reading unstructured input a human wrote in a hurry, working out which of the forty exceptions this is, deciding whether this case is even the kind of case the process covers. That slice is genuinely hard and genuinely new and nothing else can do it. It is also, in most back offices, the minority of the work by volume.\n\nAn honest vendor will tell you which fifth of the system is AI. The incentive runs hard the other way, because the specification is consulting revenue that ends and the agent is a subscription that does not.\n\nA company does not run on tasks. It runs on **who signed off**.\n\nEvery consequential action in an organisation has a name attached to it. Who approved the credit. Who authorised the discount. Who released the shipment, waived the fee, closed the ticket, accepted the risk. This looks like bureaucracy from the outside and it is not. It is the load-bearing structure. It is how the company learns from being wrong, how it prices its own risk, how it satisfies an auditor, and how the person who has to answer for an outcome is able to answer for it.\n\nAn autonomous agent has no name.\n\n*\"The system did it\"* is not an answer that any auditor, regulator, insurer, board, enterprise\ncustomer or wronged member of the public has ever accepted, and there is no version of model progress\nthat changes this. It is not a technical limitation. It is what accountability *is*: a person who\ncan be asked, and who could have chosen otherwise.\n\nThis is the actual ceiling on the category. Not hallucination. Not accuracy. Not context windows.\nThe reason so many pilots die at exactly the boundary between reading and writing is that read-only\nnever has to answer the question, and the first write does. You get through six weeks of enthusiastic\nevaluation and then you meet the person whose signature is on the control, and they ask the one thing\nno accuracy number answers: **when this is wrong at three in the morning, whose name is on it?**\n\nAlmost everyone in this market is optimising toward zero humans in the loop. It is the implicit objective function of the whole category — every roadmap is a march toward the agent needing you less.\n\nI think that is the wrong objective function, and not for sentimental reasons.\n\nThe right one is: **collapse the cost of the human decision until the loop is close to free.**\n\nConsider what the human is doing in each case. Doing a short-shipment credit by hand means opening the ERP, finding the order, finding the proof of delivery, comparing two documents, computing a number, navigating four screens and typing it in. Approving a proposed one means reading a diff and pressing a key. Those are not the same activity and they are not within an order of magnitude of each other in cost.\n\nSo the arithmetic of the whole category is this. Moving from *a human does the work* to *a\nhuman approves the work* is where nearly all of the available gain lives. Moving from *approves*\nto *does not look at all* is the last sliver — and you pay for it with the only artefact anybody in\nthe building actually wanted, which is a name on the decision.\n\nFull autonomy is the least valuable increment of automation and by far the most expensive\none to defend.\n\nThe companies racing to remove the human are spending their hardest engineering on the cheapest remaining percentage points, and they are spending it on the exact axis that makes the buyer with signing authority say no.\n\nAccept those two claims and the design is close to forced. This is not a philosophy, it is a build specification, and every item on it is a thing we will not ship without.\n\nNo API key, no database connection, no authenticated session, no browser cookie. The model reads a\nbounded context and emits a *proposal* into a queue. A separate execution layer, which contains no\nmodel, decides whether to perform it.\n\nThis is not a security nicety bolted on at the end. It is the thing that makes every other property on this list possible. It also collapses prompt injection from a catastrophe into an annoyance: the worst outcome of a poisoned input is a bad proposal, and a bad proposal is a thing a person rejects in one second and a rejection taxonomy records forever.\n\nNot free text. Not an arbitrary tool call with arbitrary arguments. A finite set of verbs that the executor is willing to perform, each with a schema.\n\n```\n{\n  \"action\":   \"issue_credit\",\n  \"order_id\": \"SO-118422\",\n  \"amount_cents\": 41800,\n  \"reason_code\": \"short_ship\",\n  \"evidence\": [\"pod:8871\", \"inv:SO-118422\"],\n  \"reversible_until\": \"2026-09-08T17:00:00Z\",\n  \"confidence\": 0.91,\n  \"abstain\": false\n}\n```\n\nIf the model wants to do something outside the vocabulary, it cannot ask for it. That is the feature. A closed vocabulary is what lets you reason about the blast radius of the system as a whole rather than about the behaviour of a model you did not train and cannot audit.\n\nThe execution layer independently re-reads live system state and refuses the proposal if the world has moved: the order already shipped, the credit already exists, the amount exceeds the policy ceiling for this action type, the record is closed, the customer is on hold. It refuses on its own authority, in ordinary code, with tests.\n\nThe model's confidence score is not an input to that decision. It is telemetry.\n\nReversibility is a property of the action type, decided in advance and written down, not a hope expressed at review time. Anything with an irreversible external side effect — money leaving the building, a message reaching a customer, a regulatory filing — either sits behind a recall window that is genuinely enforced, or it stays in human hands permanently.\n\nA system that must always produce an answer is a system that will invent one. *\"I do not know,\nhere is what is missing, here is who should look at it\"* has to be a first-class, structured,\nrewarded outcome — not a fallback, not an error branch.\n\nPut the abstention rate on the dashboard next to the accuracy rate. If you only ever measure and celebrate accuracy, you have quietly trained the entire system, and the team operating it, to bluff.\n\nNot a global slider labelled \"how much do you trust the AI\". Correcting a delivery address and issuing a four-hundred-dollar credit are different risks and get different ladders.\n\n| Rung | What happens | \n|---|---|\n| **L0 — shadow** | Proposes into the dark. Nobody acts on it. You are measuring, not deploying. | \n| **L1 — draft** | Proposal becomes a draft a human edits and sends. | \n| **L2 — queue** | Human approves or rejects each one, with a reason code. | \n| **L3 — auto + recall** | Executes immediately, reversible inside a stated window, reviewed after. | \n| **L4 — autonomous** | Executes, sampled for audit. Very few action types ever get here, and some never will by design. | \n\nGraduation between rungs is a number agreed before the project starts — a volume of decisions and a rejection rate below a stated ceiling — not a judgement made in a steering meeting by whoever is most enthusiastic that quarter.\n\nEvery proposal, the evidence it rested on, the decision, the person, the timestamp, the resulting\nstate change, and the rollback path. That record is what you hand an auditor. It is what makes\nautomating the *second* process a twenty-minute conversation instead of a six-week evaluation. It\nis the only reason anyone gives you write access to a third.\n\nThe AI is not the product. The AI is what fills in the ledger.\n\n**One: the well-funded autonomous-agent companies converge on the approval queue.** Not because\nthey want to — it is off-narrative and it makes the demo less impressive — but because that is where\nenterprise purchasing actually happens. The person with signing authority is the person whose name is\non the control. They are not buying autonomy. They are buying leverage they can defend in a meeting\nafter something goes wrong. Every company in this category eventually meets that buyer.\n\n**Two: whoever owns the ledger owns the account.** The agent is swappable and getting more\nswappable every month; the lock-in on the reasoning layer is close to zero and everyone in this market\nquietly knows it. The structured record of every operational decision a company made for three years,\nwith evidence and outcomes attached, is not swappable. It is also, incidentally, the highest-quality\ntraining and evaluation data that will ever exist for that company's specific work — which means the\nledger is not just the moat, it is the flywheel.\n\nIf you run operations somewhere and you want to start this without hiring anybody:\n\nThat last one is not defensive PR. The people who know where the forty exceptions live are the people whose cooperation determines whether any of this works, and the fastest way to lose them is to leave them guessing where the line is. Tell them where it is. They will hand you the specification you came for.\n\nSupport automation, AI phone agents, n8n back-office work, and the engineering loop itself — always behind a gate you control.", "url": "https://wpnews.pro/news/company-does-not-have-processes-it-has-habits", "canonical_source": "https://ai.levelbrook.com/playbook/you-do-not-have-processes/", "published_at": "2026-09-07 16:06:21+00:00", "updated_at": "2026-09-07 16:26:54.230447+00:00", "lang": "en", "topics": ["ai-agents", "ai-products", "ai-startups"], "entities": [], "alternates": {"html": "https://wpnews.pro/news/company-does-not-have-processes-it-has-habits", "markdown": "https://wpnews.pro/news/company-does-not-have-processes-it-has-habits.md", "text": "https://wpnews.pro/news/company-does-not-have-processes-it-has-habits.txt", "jsonld": "https://wpnews.pro/news/company-does-not-have-processes-it-has-habits.jsonld"}}