The AI that watches how work actually happens - then politely points out where it could happen faster.
Section 1 · who they are, right now
Eleven o'clock on a Tuesday. A Fortune 500 finance team is on a Zoom that has been running for forty-seven minutes. Somebody is sharing a screenshot of a process diagram drawn in 2019. Somebody else just said the words "tribal knowledge" without irony. The transformation consultant - billed at $4,200 a day, currently on slide 38 of 91 - is nodding.
This is the room Klarity is trying to empty out.
The San Francisco company, now about 150 people strong, sells an enterprise AI platform that does something disarmingly simple. It watches. It watches Zoom recordings, screen activity, and the documents people pass around. Then it writes down what it saw. Quietly. Within minutes. In a format anyone can actually use.
Section 2 · the original heresy
Every large company has, somewhere inside it, a dark matter of knowledge: how the AR team actually closes the books, which spreadsheet really runs procurement, what the workaround is when the system says no. None of it is written down. None of it survives a resignation. All of it is, technically, billable to a consulting firm.
The orthodox fix is to send in humans with clipboards, run six-month "current state" interviews, and produce a report that is obsolete the day it's delivered. Klarity's bet was that this entire ritual could be replaced with software.
CAPTION - A management consultant explaining what a management consultant does. Photographed candidly. Visible boredom is real.
The bet was not obvious in 2017. It is closer to obvious now.
Section 3 · how it started
Andrew Antos was a Harvard-trained lawyer who claimed, with a straight face, that reviewing thousands of contracts was fun. Nischal Nadhamuni was an MIT computer science graduate who, on hearing this, did the only sensible thing: he sketched a neural-network architecture for it. They met in an MIT class. They argued. Then they founded a company.
They called it Klarity Law. The name was a tell. The original wedge was contract review for lawyers - an audience the founders soon discovered has the budgetary enthusiasm of a librarian and the procurement cycle of a sovereign government. By 2020 the pivot was underway. The product was the same; the buyer was different. Klarity would automate document review, but for the accounting function, where the documents are arguably duller and the urgency is, blessedly, much higher.
The lesson was not subtle. Useful technology applied to a customer who cannot move quickly is just an art project. Klarity stopped doing art projects.
Section 4 · the product, plainly
The current platform has three phases, and they are named with admirable bluntness: Discover, Structure, Improve.
An AI Companion passively watches workflows across the apps employees already use. AI Interviews then ask the questions a good consultant would, but at conversational scale. There are no integrations required - a fact Klarity repeats often, because anyone who has tried to plumb data into an enterprise warehouse knows what that sentence is actually worth.
The Process Index turns the captured flow of work into a navigable library. Every SOP. Every handoff. Every quiet workaround the new hire never gets told about.
Advisor projects ROI on proposed changes. Signals flags anomalies in real time - the equivalent, in operating terms, of a smoke detector for processes that have started to drift.
Section 5 · proof of work
It is one thing to claim that AI can replace a transformation consultant. It is another to hand the keys to the CFO of DoorDash. Klarity has done the second one - repeatedly. The customer roster reads like the standard slide every enterprise AI company aspires to but only a handful ever earn.
CAPTION - The shape of a company that took its time before it took the money. The 2024 bar is the one investors point at.
Section 6 · why they bother
The company's own line, taped to its About page: "We're building the tools that let change agents change things." It is the kind of sentence that would be insufferable if it weren't, on closer inspection, accurate. Most enterprise software claims to enable transformation. Klarity's claim is narrower and harder to wriggle out of - it wants the person inside the company who is trying to actually change something to finally have a tool that doesn't make them file a ticket first.
The internal values are four words: Velocity, Agency, Care, Energy. Standard issue, except for the ordering. Velocity goes first. Slowness is the enemy. This is a company that has watched enough enterprise sales cycles to know that the only sustainable competitive advantage is being faster than the people who could otherwise stop you.
Section 7 · the bigger picture
The line CFOs hate to talk about in public is the G&A line. It is the cost of running the business that does not, technically, run the business. Finance, IT, HR, legal, ops. It grows linearly with headcount and almost never gets cheaper without somebody losing their job. Klarity's pitch is that the line can be bent - that the curve can be flattened by software that knows what people in those functions are actually doing, before they themselves can fully articulate it.
If that pitch is right, it implies something larger. It implies that the next decade of enterprise software is not about replacing humans with chatbots, which everyone is bored of discussing, but about giving the operators of large companies a kind of x-ray vision they have never had. Not a dashboard. Not a report. A live model of how the business actually behaves, updated continuously, defensible to auditors, and useful by Wednesday afternoon.
The funders agree. Nat Friedman and Daniel Gross - the GitHub-and-YC alumni now known to a generation of founders as NFDG - led the Series B personally. Scale Venture Partners joined. Tola, Picus, Invus, and Y Combinator wrote follow-on cheques. The dealflow at NFDG is competitive in the way that an Olympic trial is competitive. Klarity made the team.
Section 8 · the close
Eleven o'clock on a Tuesday. The same finance team. The same Zoom. The consultant is gone. Slide 38 of 91 has been replaced by a process index that updates itself overnight, drafted by a piece of software that watched the team work last week and quietly wrote down what it saw.
The meeting is shorter now. The decisions are faster. The tribal knowledge is no longer tribal. Somebody on the call is, possibly for the first time, looking at a complete picture of how the company actually operates - and they can edit it.
That is Klarity's pitch. That is, increasingly, Klarity's product. And that, eventually, will be a meeting nobody dreads.