DevRev hits $1.15B valuation in August 2024 1,000+ SaaS customers on AgentOS Founded by Nutanix alumni Dheeraj Pandey and Manoj Agarwal $100.8M Series A led by Khosla Ventures and Mayfield Knowledge graph + AI agents = new CRM stack Palo Alto - Bangalore - 840 employees and counting DevRev hits $1.15B valuation in August 2024 1,000+ SaaS customers on AgentOS Founded by Nutanix alumni Dheeraj Pandey and Manoj Agarwal $100.8M Series A led by Khosla Ventures and Mayfield Knowledge graph + AI agents = new CRM stack Palo Alto - Bangalore - 840 employees and counting
DevRev logo
Company Dossier Vol. 04 / AI-Native CRM

DevRev

A four-year-old SaaS unicorn arguing the loudest, with the most receipts, that CRM as we know it is over. AgentOS, a knowledge graph, and a tight bet on agents over tabs.

PORTRAIT: A logo that looks like two arrows arguing politely. One points at the customer. The other points at the engineer. They agreed to share a database in the middle.
SHARE - Twitter LinkedIn Facebook Instagram
A YesPress Profile / Palo Alto Bureau / Filed May 2026

"Customer Relationship Management" was always a slightly polite name for a spreadsheet. DevRev wants to replace it with something that thinks.

The Tuesday Standup

It is Tuesday morning in Palo Alto, and somewhere on Hamilton Avenue an AI agent is closing a support ticket that a human has not yet read. The customer wrote in at 3:14 a.m. Mumbai time, polite and a little frustrated. By 3:14:08 their question had been routed, parsed, matched against the product's full history, and answered in language that sounded almost suspiciously human. The agent did not pat itself on the back. It moved on to the next thread. There are 840 employees at this company. None of them did that work. All of them get credit for it.

This is DevRev now. Not a startup with a clever slide, not a thesis being workshopped on Notion. A four-year-old unicorn that has already shipped what other vendors are still describing on a stage. The pitch is not subtle. The old way of running a SaaS business - one tool for sales, one for support, one for product, glued together by exhausted operations teams - is, in DevRev's telling, a museum exhibit. They would like to be the wrecking ball. Politely, of course. With excellent typography.

CRM was always a spreadsheet wearing a tie. DevRev took the tie off and gave it a brain.- The thesis, abridged

The Problem Nobody Wanted to Name

Every modern software company runs on the same dirty secret. The customer is in one system. The product is in another. The roadmap is in a third. When a support ticket comes in about a broken checkout flow, it bounces between four tools, three teams, two time zones, and at least one engineer who has to ask, again, what version the customer is on. The bounce costs days. Sometimes weeks. Often the customer leaves.

DevRev's co-founders had spent a decade at Nutanix watching that bounce happen at enterprise scale. Dheeraj Pandey, who took Nutanix public and built it into a sixteen-billion-dollar story, did not need another company. He had a board seat at Adobe and a comfortable runway to do whatever he wanted. Manoj Agarwal, his longtime SVP of engineering, did not need one either. They started one anyway. In late 2020, in a year when most people were learning to bake bread, they were writing a manifesto about why Salesforce was built for the wrong century.

CAPTION - FILE 01: A founder who has already won once is the most dangerous kind. He has nothing to prove and a list of things he would like to fix.

The Bet

The bet was simple to say and reckless to ship. Build a single system, not seven. Put a real-time knowledge graph at the center. Make every customer message, every product feature, every line of code, every roadmap item a node in the same graph. Then drop AI agents on top of it - not chatbots stapled to a sidebar, but agents that can read the whole map and act on it. Call the bundle AgentOS. Sell it to support, product, and growth teams who had quietly accepted that their tools would never talk to each other.

Investors wrote the seed check before the product had a name. Khosla Ventures and Mayfield put in fifty million dollars on the strength of two founders and a deck. The product would arrive when it arrived. Engineering took three years of head-down work before DevRev felt confident enough to throw a launch party. By the time it did, AgentOS was not a demo. It was running in production at companies most readers will recognize.

Most startups raise to find a problem. DevRev raised because the founders had already found one, twice.- An investor, paraphrased

What AgentOS Actually Is

Strip the marketing off and the platform breaks into three pieces. Support handles tickets, conversations, and deflection - AI triages incoming queries, drafts replies, escalates only what humans must touch. Build is the product workspace, where engineering tickets, customer feedback, and roadmap items live in the same graph rather than three different tools. Grow is the lifecycle layer, watching how customers use the product and quietly suggesting which ones are about to churn, expand, or upgrade. The trick is not that any one piece is revolutionary. The trick is that they share a brain.

That brain is the knowledge graph. In a traditional CRM, a customer is a row in a table. In DevRev, a customer is a constellation - a node connected to the features they use, the tickets they have filed, the engineers who wrote those features, the version of the product they are on, the conversations they have had with sales. Ask any question of that graph and you get an answer rooted in the whole story, not the slice of it that happened to be entered into a form. Agents query the graph. Humans query the graph. The graph just sits there and remembers.

BUILT WITH

TypeScript and React on the front. Vitess, DuckDB, Kubernetes, Argo on the back. Claude doing some of the thinking.

SHIPPED FROM

Two headquarters - Palo Alto and Bangalore. Video over docs internally. Slack over meetings. Standups over status updates.

SOLD TO

Support, product, and growth teams at SaaS and tech companies. Roughly 1,000 paying customers, and counting more quietly than the marketing would suggest.

NAMED BECAUSE

"DevRev" is Developer + Revenue. The two halves of every software business that, historically, never sat next to each other.

A Four-Year Sprint

MILESTONES, COMPRESSED
2020
The founders leave NutanixPandey and Agarwal raise a $50M seed from Khosla and Mayfield. The product does not exist. The thesis does.
2022
First public walkthroughsDevRev shows the platform externally. The pitch lands harder than expected with engineering-led companies.
2023
Customer count rocketsThe platform crosses what insiders describe as "the embarrassing number" of paying customers in twelve months.
2024
Unicorn status$100.8M Series A. $1.15B valuation. One of the largest Series A rounds in SaaS that year.
2025
AgentOS expandsFull product walkthroughs ship. Pandey publicly argues that the real AI moat is integration, not model size.
2026
Now840 employees. Two HQs. A platform sitting between customer and code at over a thousand companies.

The Proof

A thesis without customers is a sermon. DevRev's customer list reads more like a who's who of companies that obsess over support latency than a logo-slide for a board meeting. Uniphore is on it. One of the world's top-five consumer banks, with a hundred million users, is on it. Some of the largest AI chip designers in the world use DevRev to keep product information and customer questions tied to the same thread. The unifying detail is not industry. It is a willingness to admit, in private, that legacy CRM was costing them quarters.

The Numbers, as Public

SOURCES: COMPANY DISCLOSURES, PRESS COVERAGE
Valuation
$1.15B
Total Funding
$150.8M
Series A
$100.8M
Customers
1,000+
Employees
~840
Years Old
5

CAPTION - READING THE LINES: A graph that flatters the company, drawn from numbers the company has flattered itself with publicly. Take the proportions seriously, the precision less so.

A thousand customers in under two years is either a market that was begging to be fixed, or a sales team that does not sleep. The answer is both.- Field note

The Mission, in Their Words

Ask DevRev what it is for and you get a sentence that sounds suspiciously like a manifesto. Bring developers and customers closer together. Erase the wall between the people who build the product and the people who use it. The wall, in most software companies, is load-bearing. Removing it requires more than software. It requires a willingness to admit that the org chart got something wrong.

Pandey's own line, repeated in interviews through 2025, is that the real AI disruption is not the model. It is the integration. Anyone can wire a chatbot to a help center. Very few teams can put every customer message, every code commit, every product decision in the same graph and let an agent reason across all of it. That is the moat. That is also, conveniently, what DevRev sells.

Five Things That Amuse

Tomorrow Morning

It is Wednesday now. The agent that closed the 3:14 a.m. ticket has closed four hundred more like it. The customer who wrote in is back, this time with a feature request, and the agent has filed it as a node in the same graph that holds the original complaint. An engineer in Bangalore will see it tomorrow, tied to a roadmap item, with the customer context already attached. There will be no toggling between tabs. There will be no asking which version. There will be no bounce.

This is what DevRev meant when it said it wanted to retire CRM. Not a press release. A Tuesday standup where nobody opens Salesforce because nobody needs to. Whether the rest of the SaaS industry follows is, as it always is, a question of momentum, sales motion, and how loud the incumbents shout. The company has the receipts. The market has the appetite. The bounce, for once, has somewhere to land.