The modern accounting firm is organised like a Roman legion. Information moves up the hierarchy. Decisions move down. Humans sit at every level as conduits — juniors prepare, seniors review, managers consolidate, partners decide. The structure is so familiar it barely registers as a structure. It is also two thousand years old.
The legion was an astonishing achievement: it coordinated tens of thousands of people using nothing but messengers, written orders and a rigid chain of command, because information had to be physically carried, summarised and judged by a person. Modern firms inherited the shape of that system and built telephones, email and Teams on top of it. Each new tool sat inside the structure rather than replacing it. The legion stayed intact. The question this series opens with is whether that is still the right shape.
What the structure is actually for
The hierarchy is not arbitrary. It does at least four things the firm genuinely needs. It moves information — each layer compresses and filters until the partner sees a pre-digested question. It allocates judgement — a junior can’t sign off accounts; a partner needn’t choose a VAT code. It trains people — the pyramid is a school, where the career path and the production line are the same thing. And it distributes risk — a partner’s signature is a chain of professional accountability that satisfies the regulators and the insurer. Any conversation about restructuring that doesn’t take all four seriously isn’t worth having.
What is changing underneath it
The hierarchy was built for a world in which moving and processing financial information required humans at every step. A modern AI system can ingest a year of transactions, invoices and payroll, reconcile them, flag anomalies, draft journals and produce first-pass working papers in minutes. It can read a lease and pull out the treatment. It can answer most of the questions a junior would have asked a senior, by reading the firm’s previous answers. This is not a prediction — it is what the technology does today in firms that use it.
The information-moving function was the pyramid’s primary function, and it is the one AI is best at absorbing. Bolt AI onto the existing legion and you get a marginally faster legion: juniors prepare faster, seniors review faster, productivity rises perhaps twenty percent. That is not nothing. But it is not what the technology is capable of.
Drowning in manual, repetitive work? Tell us the task and we’ll show you what to automate.
The deeper question
The harder question is what the firm looks like if you take the information-moving function out of the human hierarchy entirely. Not “AI assists the junior” but “AI does the work the junior was doing”. The pyramid doesn’t collapse, because judgement allocation, training and risk distribution are still needed — but they are no longer welded to information movement, so they can be reorganised.
Judgement allocation becomes a question of which decisions actually require human judgement, and at what level. Training becomes the hardest problem of all: if juniors aren’t learning by preparing files, what are they learning by, and how do they qualify? Risk distribution becomes a question of how regulatory accountability survives when significant work is done by systems. These are not insurmountable — they are the actual problems, and they are firm problems, not technology problems, which is precisely why no vendor is helping firms solve them.
A closing observation
The legion didn’t lose its dominance because someone invented a better hierarchy. It lost it because the world changed in ways the hierarchy wasn’t designed for. It was still extraordinarily good at what it did; the thing it did was just no longer the thing that mattered most. The accounting-firm pyramid is extraordinarily good at what it does, too. The question is whether what it does is still the thing that matters most — and that is a question every managing partner should be asking out loud, with their leadership team, for at least one afternoon.
Part 1 of The Self-Improving Firm. Daniel Lawrence is the CEO and co-founder of Bots For That and creator of the automation operating system.
