01 / WHO THEY ARE NOWThe reply that doesn't come from a human.
It's a Tuesday at 9:14pm. Someone in Ohio downloads a whitepaper and forgets about it by Wednesday. By Friday, they have three thoughtful emails, a follow-up question, and a calendar invite. None of it came from a person.
That is the everyday product of Conversica, headquartered in San Francisco and quietly renamed The Conversation Company in August 2025. Its software runs what the company calls Revenue Digital Assistants - AI workers that engage prospects, customers and partners over email, SMS and chat, in a two-way back-and-forth that reads, frankly, like a real coworker who has read your last email and remembered it.
The team is around 200 people. Annual recurring revenue sits near $24.6M. Total raised: roughly $148M across five named rounds. The platform has handled, by the company's count, more than a billion conversations. None of these numbers are particularly loud by 2026 AI standards. The product is.
02 / THE PROBLEMEvery CRM is a graveyard.
Walk into the average B2B sales organization and you will find a CRM packed with leads no one has touched in months. Marketing keeps pouring water into the top of the funnel. The sales team, reasonably, only chases the leads that already look hot. The rest - the curious, the maybe-laters, the wrong-month buyers - quietly rot.
Estimates vary, but the industry consensus is grim: somewhere north of two thirds of marketing leads never receive a meaningful follow-up. That's not a CRM problem. It's a human attention problem. People have other things to do, like talk to humans who already raised their hand.
Conversica's founding bet, in retrospect, looks obvious. If the bottleneck is patience and persistence, hand that part to software. Let humans do the closing.
03 / THE BETA car-dealership CRM with a polite robot inside.
The company began life in 2007 in Bellingham, Washington, under a name nobody but its earliest customers remembers: AutoFerret.com. Founder Ben Brigham was selling CRM tools to car dealerships - a market that, by 2010, would be quietly thrilled to outsource its follow-up emails to a robot.
By the early 2010s the company had pivoted from CRM to lead engagement. It built an email-only assistant nicknamed "Ava." In 2014 the corporate name followed her: AVA.ai, then Conversica. By 2015, Foster City, California; by 2025, San Francisco proper.
The technical bet was almost contrarian: don't try to make a chatbot that pretends to be human in a thirty-second pop-up. Make an email assistant patient enough to handle a six-week nurture cycle, and good enough at language that nobody on the receiving end knows it isn't Sarah from sales.
04 / THE PRODUCTRevenue Digital Assistants, plural.
Conversica's modern platform sells what it calls Revenue Digital Assistants - RDAs, if you must - that come pre-trained for specific jobs along the revenue lifecycle. There's an assistant for inbound lead qualification. One for outbound demand generation. One for customer success that politely asks if your trial seat is still useful and whether you'd like to talk renewal. One for partner enablement.
What's under the hood
Until recently, the answer was a combination of natural language understanding, classification models, and a heavily curated library of conversational paths. Since the LLM wave, the company has woven generative models into the platform - alongside retrieval-augmented generation, guardrails, and a multi-model approach that picks the right engine for the right turn of the conversation. It's less "we slapped GPT on top" and more "we already had the orchestration; now the brains are better."
05 / THE PROOFFive rounds. One stubborn idea.
Investors have written checks every two to three years since 2013. The composition of that cap table reads less like a hype cycle and more like a slow build: a growth-equity firm (Providence Strategic Growth) returning for two consecutive rounds, then a Morgan Stanley-managed fund stepping in at Series E. The kind of money that wants product-market fit, not poster placement at a conference.
FUNDING / NAMED ROUNDS
Customers cluster in places where the cost of an unworked lead is genuinely measurable: automotive groups (the original tribe), financial services, telecom, education, technology and enterprise SaaS. The integrations list reads like a tour of where revenue actually lives - Salesforce, HubSpot, Marketo, Eloqua, Microsoft Dynamics, plus dealer-tech mainstays like VinSolutions, DealerSocket and CDK. Old roots, new branches.
06 / THE MISSIONUnlock the knowledge nobody can answer fast enough.
The August 2025 rebrand to The Conversation Company is not, mercifully, a logo swap. It is a thesis. CEO Mike Gilson's pitch, paraphrased: every enterprise is sitting on a pile of knowledge - product docs, pricing, policies, account histories - that customers want answers from in the language of a conversation. LLMs make those conversations scalable. Conversica's job is to wrap them in workflow, integration, brand voice and guardrails so the answers actually drive revenue instead of just dispensing trivia.
It is a slightly less glamorous mission than "AGI by Tuesday." It is also, on inspection, more useful.
07 / WHY IT MATTERS TOMORROWThe autonomous coworker question.
The interesting fight in enterprise AI for the next several years is not whether software can answer a question. It can. The fight is whether software can start a conversation - politely, persistently, at scale - and carry it through to an outcome a CFO can point at on a dashboard.
Conversica's seventeen-year head start on that specific problem is not nothing. Most LLM-native challengers are still learning what every long-suffering Conversica engineer already knows: that the hard part of a sales conversation is not the writing. It is the cadence, the handoff, the silence between message four and message five, the moment a buyer says actually, ask me again in March.
08 / BACK TO TUESDAYThe reply, revisited.
Return, briefly, to that whitepaper download on Tuesday at 9:14pm. In the old version of this story, the lead would die in a CRM next to twenty thousand others. A salesperson would have triaged it on Wednesday morning, decided it didn't look ready, and moved on. By March, the prospect would have bought from a competitor.
In the Conversica version, the lead gets a note within minutes. A real question on Thursday. A polite nudge a week later, when the human goes quiet. By the time a salesperson walks into the meeting on Friday, the prospect has been listened to, qualified and warmed up by something that, on paper, should not be able to do any of that.
That is the conversation Conversica has been having - patiently, two ways, a billion times over - since before most of its competitors had a name. The robot, it turns out, was the patient one all along.