Most AI agents get added to call centers for one reason — to cut down on how many humans you need.
But a startup that just raised $30M in Series A funding did the opposite. Instead of asking "how do we handle fewer calls," they asked "which conversations actually turned into deals?"
Here's what everyone assumes
Most AI support and CS agents today are judged on one thing: how fast they can resolve a query without a human. Chatbots and voice AI already handle up to 40% of inquiries. Gartner predicts that by 2029, agentic AI will independently handle 80% of routine service inquiries, cutting operational costs by 30%.
Under this logic, AI is a device for avoiding humans. Most success metrics are cost metrics — resolution time, ticket volume.
But the numbers went the other way
Dr. Dvir Ginzburg, who holds a PhD in geometric deep learning from Tel Aviv University and worked on recommendation systems at Microsoft, founded Encore AI (formerly Insait) in 2022 — and flipped the formula. As the CEO put it bluntly: "Today's enterprise AI industry is optimized for cost reduction. Encore was built for the other side." While most AI agents are designed to deflect calls and reduce human contact, Encore AI says it was built to increase revenue instead.
The method is a patented approach called Interaction Mining. It ingests millions of call recordings, emails, texts, CRM records, and landing-page data, breaks conversations into stages, and identifies which moves actually advanced a deal. It extracts the repeated behaviors of an organization's top performers and turns them into either a real-time assistant called "Wingman" or a fully autonomous agent called "Autopilot."
The signal that this is working shows up in the numbers. ARR is up more than 5x in the 18 months since the seed round. One lending client reportedly saw a 10x return on investment within months. The round was led by Team8, with Planven and The Garage participating — and notably, some of the investors were banks and insurers that started out as customers before deciding to put money in.
Why financial services first
Most of Encore AI's 40-plus enterprise customers are financial institutions. It seems counterintuitive to target the most heavily regulated industry first, but that's exactly what turned into a moat.
Compliance requirements keep most AI startups out of regulated industries entirely. Financial services also became the lead vertical because of how often hidden revenue opportunities slip past unnoticed — a mortgage call that turns into a remodeling question, a business customer asking about cash flow. One financial institution reportedly processed 200,000 calls and saw an 81% increase in conversion, saving 1,300 hours in the process. The company says deployment can happen within weeks, compared to the months typical enterprise AI rollouts take.
How to try this without Encore AI
You don't need Encore AI's platform to borrow the "interaction mining" mindset.
- Collect your winning and losing conversations
Pull 20 closed-won and 20 closed-lost calls or emails from the last three months. That's enough to start. - Break them into stages and tag them
Split each conversation into opening, needs discovery, objection handling, closing — and mark what repeats only in the wins. - Start with real-time assist, not full automation
A hint system for reps ("this question worked well at this point") is lower-risk than full automation. - Measure by revenue, not efficiency
Skip resolution time and ticket volume. Compare conversion rate and deal size between the group using this approach and the group that isn't. - Check compliance first
Using call recordings and customer data to train AI has different consent and retention rules by industry. Loop in legal before you pilot.
Go deeper
Encore AI raises $30M to build AI agents that learn from customer calls The original story — founding background and the full funding round context techcrunch.com
Encore AI raises $30 million Series A to teach AI agents how top employees work A deep dive on founder Dvir Ginzburg's background and the Interaction Mining tech calcalistech.com
Encore AI's official site Details on the Wingman and Autopilot modes plus customer case studies by industry gainencore.ai
Decoding Conversation Intelligence in Sales and Marketing Gong's take on conversation intelligence — useful for comparing against Encore AI's approach gong.io
Encore AI Raises $30M Series A To Deploy Revenue Agents A CX-industry take on what this funding round signals cmswire.com
Encore AI Raises $30M for Revenue-Generating Voice Agents Industry trend data, including the Gartner forecast, alongside the funding analysis enterprisedna.co
