The CRM is the strangest product in business software. Nearly every company buys one. Almost nobody enjoys using one. Salespeople treat it as homework, managers treat it as a lie detector that lies, and the data inside it is stale roughly from the moment someone types it in. Attio, a company founded in London in 2019, looked at this quarter-century-old bargain and decided the problem was not the salespeople. It was the software.
Attio builds what it calls an AI-native CRM - a customer relationship system designed from its data layer upward for a world where software agents, not just humans, do the work of finding, researching, and closing customers. Its homepage describes the product as "the CRM for agentic revenue" and, more memorably, "the intelligent system that never sleeps." More than 30,000 active customers now run on it, and the roster reads like an index fund of the AI boom: OpenAI, Cursor, Harvey, Linear, Notion, Ramp, Granola, Modal, Replicate, and TaskRabbit among them.
That customer list is the tell. The companies that grew up in the last five years - selling by usage, growing through their product rather than a phone bank - kept discovering that the CRMs built in 1999 and 2006 could not describe their business. Attio built the CRM that could.
The unfashionable years
Attio was founded by Nicolas Sharp, its chief executive, and Alexander Christie, its chief technology officer. What the two did next was almost aggressively boring: they spent years on data architecture before chasing attention. While competitors shipped features, Attio built a relational data model - a system where a company is not forced into a rigid grid of "Contacts" and "Accounts," but can define its own objects, link them however its business actually works, and trust that everything stays synced.
That patience became the product. In Attio, a developer-tools startup can model workspaces, users, and deals as first-class objects. An investor can model funds, LPs, and portfolio companies. The interface earned comparisons to Notion and Linear - the design-forward tools its customers already loved - but the deeper appeal was structural. As Christie put it, "Today's go-to-market builders expect platforms they can shape to fit their vision, not rigid systems."
The second foundational decision was to kill data entry. Attio syncs email and calendar automatically - about 15 million emails a day now - and enriches records from live sources, so the CRM fills itself. This addressed the original sin of the category: a CRM is only as good as what humans remember to type into it, and humans remember very little on a Friday afternoon.
"CRM is one of the most important categories in B2B, but it's been stuck in the past." Nicolas Sharp, co-founder and CEO
From filing cabinet to coworker
If flexible data was act one, agents are act two. Over 2025 and 2026, Attio shipped a run of AI features that changed what the product fundamentally is. A Research Agent pulls live web context onto records, doing in seconds the prospect homework that used to eat a rep's morning. Call Intelligence records meetings and writes the notes straight onto the deal. AI Attributes fill fields against rubrics the customer writes - "score this lead's fit from 1 to 10, here's what good looks like" - across thousands of records at once.
In February 2026 came Ask Attio, a natural-language layer that lets anyone query or update the CRM conversationally, in the app or from Slack. An MCP server - the open protocol that lets AI assistants use external tools - connected Attio to ChatGPT, Claude, and Raycast, and now handles some 2.6 million calls a month. The CRM became something you talk to, and something other AIs talk to.
Then, in June 2026, Attio rebuilt its Workflows product from the ground up around agents. Users describe an automation in plain language and Attio assembles it. Custom agents run over what the company calls Universal Context - the full connected record of emails, calls, notes, product usage, and external tools. Workflows can execute JavaScript mid-run for teams who want precision, and AI is billed on actual token usage rather than flat credits. The pitch, in Sharp's telling, is a system that thinks, researches, routes, and follows up across the entire revenue motion. Attio reports 76,000 active customer agents inside the product - a workforce that files no expense reports and takes no holidays.
Point Nine seeded it, Redpoint scaled it, and GV wrote the cheque that called it a platform shift.
The money agrees
Attio's funding history tracks its thesis gaining believers. Point Nine Capital led an early seed round of roughly $7.7 million alongside Headline and Passion Capital. In March 2023, Redpoint Ventures led a $23.5 million Series A with Balderton Capital and Point Nine participating. Then in August 2025 came the round that reframed the company: a $52 million Series B led by GV, Alphabet's venture arm, bringing total funding to $116 million.
GV's reasoning was blunt. "It's been 25 years since the last major platform shift in CRM," said Michael McBride, the GV general partner and former GitLab chief revenue officer who joined Attio's board. "Attio has the AI-native architecture to define the future." At the time of the raise, Attio reported more than 5,000 paying customers and said it was on track to quadruple annual recurring revenue that year. The company operates from London and San Francisco.
The business of not being Salesforce
Attio's business model is as deliberately modern as its architecture. It is product-led: a free plan covers three seats with no time limit, paid tiers run $35 per user per month for Plus and $79 for Pro on annual billing, and there are no onboarding fees anywhere in the lineup. Compare that with the enterprise CRM tradition of six-month implementations and certified consultants, and the strategy is legible - Attio wants a founder to set it up on a Tuesday night, alone, and be running workflows by Wednesday. The AI layer adds a usage-based stream, with agent work priced by the tokens it actually consumes.
| Question | Legacy CRM | Attio |
|---|---|---|
| Data model | Fixed contacts-and-accounts schema | Custom objects, relational, shape it yourself |
| Data entry | Humans type it in, eventually | Auto-synced email, calendar, enrichment |
| AI | Bolted on top | Native to the data layer; agents run workflows |
| Getting started | Implementation partners, onboarding fees | Free for 3 seats, self-serve, no fees |
None of this makes the incumbents easy prey. Salesforce remains the system of record for the enterprise, with a depth of compliance tooling and ecosystem that a seven-year-old company cannot match. HubSpot bundles marketing, service, and content tools that Attio does not attempt, and independent reviewers note HubSpot's native automation library still covers more ground. Attio's wager is that this is the wrong axis of competition - that the next decade's category winners will be decided by who models modern go-to-market motions best and who puts agents to work first, not by who has the longest feature list. Companies like Railway, Modal, and Granola switched precisely because the old schema could not describe usage-based, product-led selling.
"When I first opened Attio, I instantly got the feeling this was the next generation of CRM." Margaret Shen, Head of Business Operations, Modal
Built by builders, for builders
A quieter strength is that Attio treats developers as a first-class audience. A public API now absorbs roughly 400 million calls a week, an SDK supports custom apps, and the MCP server makes the CRM a peer in the AI tool ecosystem rather than a silo. Integrations reach into Slack, Notion, Clay, and ChatGPT. For technical founders - a large share of the customer base - a CRM you can program is not a nice-to-have; it is the difference between a tool and a platform.
Where the market goes from here is the interesting question. The CRM category is worth tens of billions of dollars a year, and every incumbent is now racing to retrofit AI onto architectures designed for a different era. Attio is running the opposite race: it started with the architecture and is sprinting to add the enterprise muscle. Its trajectory - 30,000 customers, a 4x revenue year, Alphabet's venture arm on the cap table - suggests the retrofit-versus-rebuild debate is far from settled.
What can you actually do with it? Model your business the way you drew it on the whiteboard. Let the system read your email so nobody updates a record by hand again. Point a research agent at your pipeline overnight. Ask your CRM a question in plain English from Slack and get an answer that is current. For two decades the CRM was a tax on selling. Attio's bet is that it can be the thing doing the selling - and 76,000 agents are already on the clock.
- 2019Nicolas Sharp and Alexander Christie found Attio in London.
- 2021Seed round of ~$7.7M led by Point Nine Capital.
- 2023$23.5M Series A led by Redpoint Ventures.
- 2025$52M Series B led by GV; 5,000+ paying customers, 4x ARR trajectory.
- 2026Ask Attio, MCP server, and agent-powered Workflows launch; 30,000+ customers.