People.ai:
the company that automated sales hygiene.

It's Monday. Somewhere, a forecast call is starting.

A regional VP at a Fortune 500 software company opens her laptop. The sales pipeline review is in twelve minutes. Ten years ago she would have spent the weekend chasing reps for updates. Today, she opens a dashboard, types one question - which deals slipped this quarter and why? - and reads the answer. The data was captured by People.ai while everyone was sleeping. The model wrote the explanation while she made coffee. It is, on balance, a strange way to run a sales floor. It is also the way most of the Fortune 500 now does it.

People.ai sits in the unglamorous middle of enterprise software. It is not a CRM. It is not a chatbot. It is the connective tissue between the work salespeople actually do - the email, the calendar invites, the call transcripts - and the system of record they are supposed to update but never do. The pitch is, at its core, almost rude in its simplicity. Stop asking humans to type. Let the machine watch.

A partial roster of the customers who decided their reps were better at selling than at data entry.

The data was always there. Someone just had to go get it.

Enterprise B2B sales runs on Salesforce - or it pretends to. The actual conversations happen in Outlook, Gmail, Zoom, Teams, Slack and on phones. The deals close in living rooms and conference halls. The CRM gets a sentence two days later, usually written in haste, often wrong, sometimes entirely missing. Revenue leaders have spent twenty years building dashboards on top of this swiss cheese. The dashboards are tidy. The data underneath is not.

Oleg Rogynskyy figured this out the hard way. Before founding People.ai he had spent years at sales-tech startups watching reps - smart, motivated, expensively-compensated reps - quit because they were spending their evenings filling in opportunity stages. The complaint was so universal that most managers had stopped hearing it. Rogynskyy heard it as a product specification.

An engineer walks into Y Combinator with a deeply unfashionable idea.

People.ai went through Y Combinator in 2016. The deck pitched what was essentially a robot SDR manager. The premise was unfashionable for several reasons. Sales tech was crowded. CRM was considered a solved problem - by everyone except the people using CRM. And the deep technical work - matching billions of emails to the right CRM record, deduplicating contacts across years of company history, parsing meeting transcripts into structured signals - was tedious in a way investors find hard to romanticize.

Rogynskyy did it anyway. Index Ventures put in $2.1M at seed. Lightspeed led the $7M Series A in 2017. Andreessen Horowitz joined the $30M Series B. ICONIQ - the firm that backs the personal capital of every tech CEO you have heard of - led the Series C at $60M. By August 2021 the company closed a $100M Series D at a $1.1B post-money valuation, co-led by Akkadian Ventures and Mubadala Capital. Total raised: $205M. None of it came from selling vision. All of it came from selling a working system that quietly devoured pipeline data.

The timeline of a SaaS company that is, at this point, older than most of the AI startups currently calling themselves SaaS companies.

What it actually does, in three boring sentences.

One. It plugs into the systems where work happens - email, calendar, Zoom, Salesforce, Microsoft - and silently captures every interaction between sellers and customers. Two. It uses proprietary AI to figure out which interaction belongs to which deal, which contact, which account, even when the rep forgot to BCC the CRM. Three. It surfaces the result back as forecasts, deal health flags, account plans, whitespace maps, and - since 2024 - as plain-English answers to questions like "which deals are slipping because the champion went dark?"

The result is unromantic and indispensable. Activity that used to live in twenty different inboxes now lives in one structured graph. Pipeline reviews stop being arguments about whose data is correct. Forecasts stop being optimism contests. The reps, for the first time in a decade, get to spend their evenings doing something other than typing.

1.8 million activities. Per hour. At one customer.

Cisco is the case study People.ai is happiest to talk about, and for good reason. What started as a 400-user pilot grew into 20,000+ sellers running on the platform globally. At peak, the system processes up to 1.8 million sales activities per hour - emails captured, meetings matched, contacts deduplicated, opportunities updated. Verizon, Okta, Zoom, Red Hat, IBM and Palo Alto Networks tell similar stories at different scales.

The shape of a company that didn't have a hockey-stick year but did have a lot of them in a row.

Unlock every revenue opportunity from every customer.

The mission statement is, by enterprise software standards, almost terse. There is no talk of empowering humans or democratizing intelligence. The aspiration is operational. If a customer interaction happened, capture it. If it can be matched to an opportunity, match it. If the match means a deal is in trouble, say so. If the deal is healthy, get out of the way.

This is, of course, the unfashionable opposite of how most enterprise AI is sold in 2026. There is no agent doing the work for the rep. There is no avatar joining the meeting. The rep is still the rep. People.ai's machine simply remembers everything the rep said and did, files it neatly, and writes the report the rep was supposed to write.

The rebrand, and the bet that comes next.

In April 2026 People.ai announced it was becoming Backstory. The thesis: dashboards are not what revenue leaders want. Answers are. The new name comes from a behavior already universal in pipeline reviews - the moment someone leans forward and asks, "what's the backstory on this deal?" Backstory the product is built to answer that question without a human stitching together five tools.

Whether the new name sticks - rebrands are a graveyard of good ideas - is the wrong question. The right question is whether the bet underneath it pays off. People.ai spent a decade building the substrate. The substrate is real. The activity graph at Cisco, Verizon and Zoom is the kind of asset that does not get displaced by a clever chat interface. The next decade is about turning that substrate into answers, and selling those answers to a market that has finally stopped pretending CRM works on its own.

Twelve minutes later, the forecast call ends early.

The VP closed her laptop at 10:08am. The pipeline review had taken eight minutes. There were no arguments about whose data was right. There was a slide. The slide listed the three deals slipping. Each had a one-line reason. Each had a recommended next step. The reps left to make calls. Somewhere in a datacenter, a model logged the meeting, matched the attendees, and updated the forecast. Nobody typed a thing. This is, on balance, a strange way to run a sales floor. It is also the future People.ai has spent ten years quietly building.

Four facts, lightly considered.