We started with one CPA's impossible box of paper.
Accord AI wasn't designed in a product roadmap. It was built to answer a specific favor — and then rebuilt, twice, because the people asking kept raising the stakes.
- Before
- Serial entrepreneur — several companies built and shipped
- Now
- Runs every engagement's quality bar personally
- james@accordai.net
A CPA asked whether the messy books could be automated.
James Dougherty had built companies before. What he hadn't done was sit with an accountant's onboarding problem: every new client arriving with years of disordered historical books, and a tax deadline that didn't care. Before any advice could be given — before the business conversation could even start — someone had to reconstruct the financials by hand.
The ask was narrow: could this be automated well enough to produce a trustworthy first set of financials? That question turned out to have a much longer answer than anyone expected.
A forensic client changed what "accurate" had to mean.
The first version shipped as software — a SaaS tool for cleaning up historical books. It worked. Then a CPA client brought in a forensic matter, and the tolerance for error collapsed. In bookkeeping, a misread figure is a correction. In a forensic engagement, it’s a question on the stand.
We chose to be a service, not another subscription.
The obvious move was to sell the forensic engine as software. James made the deliberate call not to. SaaS fatigue is real — nobody in a firm wants a tenth login and a self-serve upload box for evidence. More importantly, human-in-the-loop auditing isn’t a nice-to-have in this space. It’s the thing that makes AI usable at all when the output has to survive scrutiny.
So Accord AI is a technology-enabled service. We own the engine, the review infrastructure, and the analysis suite — and a US-based team reviews every extraction before it enters the record. You get the throughput of automation with a named human accountable for the ledger.
Finding the fact pattern became the obsession.
Somewhere in the work, the technical problem stopped being the interesting part. What holds our attention now is the moment a case’s shape appears in the numbers — the payee that surfaces in three accounts, the round-number cadence that shouldn’t exist. Chasing that is why the quality bar is set where it is, and why it isn’t negotiable.
Three commitments the work is built on.
A human signs off
Low-confidence fields never enter the record silently. A US-based reviewer confirms them, and the decision is persisted with attribution.
Numbers come from tools, not models
Detection and reconciliation are deterministic functions. Run them twice on the same evidence and you get the same result — which is what defensibility requires.
The engagement stays consultative
We scope around the case, not around seats. You talk to the people doing the work, not to a support queue.
Bring us your hardest case.
We'll process your pilot at a steep discount and you'll get back a structured, source-linked database.