For firms that used Botkeeper
In February 2026, Botkeeper shut down abruptly after eleven years and nearly $90 million raised; Xendoo acquired its platform weeks later. Firms mid-close learned their vendor was gone from a news article. The honest lesson is not “avoid AI bookkeeping” — it is that vendor risk is now part of vendor selection. Here is the checklist we would apply to anyone, including us.
Botkeeper customers had weeks to figure out their exit. Whatever you adopt next, test the export before you need it: full transaction history, categorizations, and documents, in a format another system can read. Nalo ships one-click export of everything — profile, transactions, receipts, categorization decisions — because a vendor that makes leaving easy is a vendor that has to keep earning your business.
The firms hurt worst by a platform shutdown are the ones whose system of record disappears with it. Tools that replace your ledger concentrate that risk; tools that run beside it do not. Nalo runs beside your books — it reads bank feeds directly and syncs QuickBooks Online read-only, so if Nalo vanished tomorrow, your books would still be exactly where they were.
Reporting on the shutdown cited 30 to 40 percent of Botkeeper’s revenue coming from roughly ten enterprise customers — when those consolidated, the company went with them. You cannot audit a private vendor’s books, but you can ask who they sell to. A product priced for hundreds of small firms decays slower than one that lives or dies with a handful of enterprise contracts. Nalo sells flat, published, per-client pricing to independent bookkeepers and small firms — the opposite concentration profile.
Automation platforms sell "98% automated." The honest question is: which 2% was wrong, and would you know? Whatever you evaluate next, demand to see the evidence behind each call, a review queue for the uncertain ones, and a close process that states what it could not verify. That is the entire design premise of Nalo: it shows the evidence behind each call, bounds what it may book on its own — small pattern-backed expenses, capped at $200, never income — and routes everything else to your review.
A migration is a bet; a parallel run is a measurement. Run any new system beside your current process on one or two clients for a month and count the disagreements before you commit. Nalo publishes exactly that as its evaluation path — the first month costs nothing, migrates nothing, and you judge the output yourself.
The same honesty, applied to us: Nalo today runs cash-basis, bank-feed books — categorization with evidence, a review queue, deterministic checks, receipts, month-close verification, and accountant exports. Full balance sheet support is in active development, and unlike Botkeeper we do not bundle a human-services team. If your firm needs either on day one, we would rather say so here than after your evaluation month.
Run Nalo in parallel on one or two clients for a month, free, and measure it against your own work. That is the whole pitch — and every question above, asked of us, has an answer you can test.
The founding partner monthSee Nalo for Firms