TLDRocket
Sign in

Encore AI raises $30M to build AI agents that learn from customer calls

TechCrunch Ram Iyer Covered by 3 sources

Encore AI just raised $30M to build voice agents that learn from your company's own sales calls. It mines what your best reps actually say and clones their pitch, jokes included.

Based on reporting by TechCrunch, Ram Iyer — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

Encore AI wants to turn every recorded sales call your company has ever made into training data. The startup, led by CEO Dvir Ginzburg, just closed a $30 million Series A led by Team8, and it's betting that the winning formula for AI customer agents isn't a generic script but a distillation of whatever your best relationship managers already do well.

The company started life in 2022 as Insait IO, building recommendation tools for financial advisers. That original engine has since morphed into something Ginzburg calls interaction mining: pulling call recordings, emails, texts, and CRM data together, breaking conversations into stages, and figuring out which moves actually pushed a deal or a support ticket toward a good outcome. The AI agents Encore then builds don't follow one playbook, they follow a composite of the ones that worked, stitched together from real transcripts. Ginzburg says the agents sometimes tell the same jokes or anecdotes a top performer would use, because the system literally learned them from the tape.

That's a different bet than most enterprise AI vendors are making. Rather than fine-tuning a model on generic customer service data, Encore treats a company's own conversational history, warts and all, as the primary asset. It also flags where a sales or support process breaks down, giving companies a diagnostic layer on top of the agent itself. More than 40 enterprise customers use the platform now, mostly banks and insurers, and Ginzburg claims annual recurring revenue has grown more than fivefold since the company's seed round less than a year and a half ago.

The obvious threat is that Salesforce, SAP, HubSpot, and Zoho already sit on mountains of customer conversation data and could build something similar without much friction. Ginzburg's counter is that these incumbents haven't architected their systems around historical dialogue as a training input, and retrofitting that would mean tearing up their existing implementation stack. Whether that moat holds for very long is genuinely unclear, but for now it's giving Encore room to run.

The new round included Planven, Lukatz, Garage, and a handful of banks and insurers, some of whom reportedly became investors only after using the product first. Encore says the cash will go toward building out U.S. sales and landing more large financial institutions, the exact customer base that seems to have convinced itself the approach works before writing the check.

My take — AI-written commentary, not fact-checked reporting

Cloning your top performer's jokes and anecdotes is a clever trick, but it's also a tell: this is really about squeezing more out of existing sales headcount, not some leap toward autonomous agents replacing humans wholesale. I'd watch the incumbent CRM giants closely here, because 'we'd have to rebuild our stack' is exactly the kind of moat that evaporates the moment a Salesforce product manager decides it's worth the sprint.

Read more about this at: TechCrunch

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.