Digits
books itself.

It is a Tuesday morning in a coworking space south of Market, and the founder of a fast-growing e-commerce company is reading her income statement. She did not request it. Nobody emailed it to her. Her accountant is asleep. The books just closed themselves overnight, somewhere inside a building on Fillmore Street, and Digits sent her the chart before her coffee arrived.

01 / Who they are nowThe startup teaching the general ledger to do its own homework

Digits is an AI-native accounting platform. That is the polite description. The more accurate one is that Digits is trying to retire the most boring weekend in every small business owner's month. Bookkeeping, reconciliations, the strange ritual of categorizing 412 Uber receipts - all of it is being quietly absorbed by software that calls itself the Agentic General Ledger.

The pitch is direct: instead of bolting AI onto a 30-year-old accounting program, Digits put AI into the foundation - the ledger itself. Transactions arrive, get classified, get reconciled, and get posted, often within seconds. Humans are kept in the loop. They are just no longer the bottleneck.

02 / The problem they sawAn entire profession waiting for the month to end

Small business finance has a peculiar shape. The data is messy, the stakes are high, and the workflow is built around a calendar event called the month-end close - a multi-day forensic exercise where humans reconcile what already happened. Most SMBs do not learn what their business did in March until somewhere in mid-April.

Jeff Seibert had seen this before. He co-founded Crashlytics, the mobile crash-reporting tool that Twitter bought in 2013. The technical lesson there was simple: developers do not want a weekly report. They want the alert the second something breaks. Why, he wondered, did accounting still operate on the opposite premise?

Caption: somewhere, a CFO is reading last quarter's numbers and pretending it is strategy.

03 / The founders' betTwo Crashlytics alums and a thesis with teeth

In 2018, Seibert and co-founder Wayne Chang - also a Crashlytics veteran - quietly began building Digits out of San Francisco. They were not, at first, especially loud about it. The product was kept private for years while the team figured out the hardest part: how to read every line of a small business's financial life without making things up.

By 2026, they had raised roughly $97.5 million across Seed, Series A, B, and C rounds, with Benchmark, GV, and SoftBank leading, and a roster of more than 70 angels that reads like a guest list at an unusually well-catered fintech conference - Aaron Levie of Box, Dick Costolo (formerly Twitter), Anthony Noto of SoFi, and Kevin Weil from OpenAI, among others.

It is the kind of cap table that does not show up unless the thesis is, as one investor reportedly put it, "structurally obvious in hindsight."

04 / The productOne ledger, several agents, zero patience for spreadsheets

Open Digits and the first thing you notice is the absence of clutter. There is no setup wizard pretending to be a friend. There is a connected bank feed, a stream of transactions, and a quiet little column that says, mostly, "categorized." If you push on it, you can see why each entry landed where it did. If you disagree, you tell it once. It learns.

The agents, briefly

AI Bookkeeping handles classification and reconciliation. Digits says it auto-books up to 95% of entries straight to the ledger. AI Bill Pay turns the pile of vendor PDFs into a drag-and-drop workflow that enriches the vendor record while it is at it. AI Invoicing drafts and sends invoices that match the company's tone instead of the template's. AI Financials produces real-time P&L, cash flow, and KPI dashboards that no longer have to wait for a person to assemble them. And underneath all of it sits the Agentic General Ledger, which is the part the accountants tend to care about.

There is a more conventional product too. Customers who prefer human hands on the books can opt into the Full-Service tier, where a dedicated accountant works alongside the AI. The pricing reflects the trade: roughly $100 a month for the software, roughly $350 and up if you want a person on call. Both layers are sitting on the same ledger - which is, of course, the point.

05 / The proofThe receipts, sorted by AI, audited by humans

Adoption numbers in private SaaS are notoriously slippery, but a few datapoints stand on their own. Approximately 78 people now build the product. The company sits at roughly $12M in estimated annual revenue, with growth steepest in the segment Digits cares most about: outsourced accounting firms running multiple SMB books in parallel.

In February 2026, Accounting Today put Digits on its Top New Products list - notable because that list is read mostly by the people whose jobs the software is, on paper, trying to automate. The accountants showed up anyway. They had read the fine print: Digits is not a replacement for accountants. It is an attempt to remove the part of accounting that no one likes.

06 / The missionMake every small business feel like it has a CFO

Seibert tends to put the company's mission in unromantic terms: bring real-time visibility and automation to every small business's finances. The romance is in the second-order effect. A founder who can see her margins on the day they shift, instead of three weeks later, makes different decisions. So does the accountant whose firm can take on twice the clients without doubling the headcount.

Digits is also unusually careful about data. The platform is built around what the team calls bank-grade security, with isolated tenant storage and audit logs on the AI's decisions - a small but important detail when the thing in the loop is a model trained on financial behavior.

Caption: the most radical thing Digits sells is the idea that you should already know.

07 / Why it matters tomorrowIf the ledger goes autonomous, so does the calendar

The interesting question Digits raises is not whether AI can post a journal entry. It already can. The interesting question is what a business looks like when month-end stops existing as an event. When the income statement is a live document. When the books are never wrong because they are never finished.

There are obvious answers - faster decisions, less reconciliation drama, fewer hours billed for janitorial work. There are less obvious ones too. If the cost of accurate financials approaches zero, an enormous tier of small businesses that currently fly blind starts flying with instruments. That is a different economy than the one we have.

08 / Back to that TuesdayThe coffee, the chart, the silence

Back in that coworking space south of Market, the founder reading her income statement does something her predecessors never had the option to do. She closes the tab. The numbers are not surprising. They are not late. They did not require a meeting. They are simply, finally, the way the rest of her software has always worked: there when she opened it.

Digits did not invent accounting. It just removed the wait. Whether that turns out to be a feature or a category remains, like most things in fintech, a question of how many more Tuesdays go by like this one.

09 / The linksWhere to look next