Encore AI raises $30 million for voice agents trained on customer calls
Encore AI just raised $30 million to turn a bank’s best call into software. The company formerly known as Insait IO is selling voice and text agents trained on real employee–customer conversations — not script trees dressed up as AI — and banks and insurers that used the product first then wrote checks into the round.
What Team8 is funding at Encore AI
TechCrunch reports a $30 million Series A led by Team8, with Planven, Lukatz, and Garage also in, plus banks and insurers that used the product first and then invested. Encore’s PR Newswire release frames the lead more broadly as Team8, Planven, and The Garage, with Lukatz and those same financial institutions alongside. The Garage also led the earlier seed and has been in since founding.
Founder and CEO Dvir Ginzburg started the company in 2022 as Insait IO, building recommendation software for financial advisers and relationship managers. The rebranded Encore AI platform now mines those interactions and ships agents that can talk to customers by voice or text, or sit beside humans and suggest the next move.
Proceeds go to U.S. sales expansion and more large financial-institution deployments.
How call-trained agents differ from script bots
Ginzburg calls the method “interaction mining.” The platform pulls call recordings, emails, and texts, ties them to CRM systems, breaks conversations into stages, and looks for what actually moved a deal or service case forward — and what stalled it.
The agent is not a generic chatbot with a bank logo. Per Ginzburg’s interview with TechCrunch, it is a package of playbooks that already worked inside that company. Agents can even reuse jokes, anecdotes, and examples relationship managers use when those patterns show up as winners in the data.
Encore’s PR push goes further on positioning: most enterprise agents optimize for deflection and cost; Encore says it was built to lift revenue — convert leads, close applications, recover balances, upsell and cross-sell — across voice, chat, IVR, and live form-fill, under regulated-industry compliance constraints. The company claims agents can go live in weeks by ingesting existing calls, chats, emails, and CRM data rather than months of greenfield tuning.
Why banks are the early beachhead
Ginzburg told TechCrunch Encore has more than 40 enterprise customers globally, the majority of them financial institutions. Banks already record calls, already live under compliance theater, and already know which relationship managers print money. They also have enough interaction volume for mining to mean something.
Investors from the customer base matter here. When a bank or insurer that already runs the product puts capital into the Series A, the diligence is partially done in production. Encore’s release also says one large lending client saw a 10x return on investment within months of implementation — a company claim, not a third-party audit, but the kind of number that shows up in every subsequent sales deck.
The competitive risk is obvious. Salesforce, SAP, Zoho, and HubSpot sit on CRM data and can ship adjacent agent features. Ginzburg’s counter to TechCrunch is that those vendors do not treat conversational history as the foundation of the agent, and that rewiring their stacks around employee conversation data would mean changing the implementation and technology stack, not flipping a feature flag.
ARR growth claims and what still needs proof
Encore says annual recurring revenue has risen more than 5x since its seed round less than 18 months ago. Ginzburg declined to disclose exact revenue or valuation to TechCrunch.
For buyers, the stake is clear. If your best closers’ habits can be mined and redeployed, the scarce asset is no longer headcount on the phones — it is the corpus of what already worked.
Watch U.S. bank logos on the customer page, and whether Encore publishes audited outcome numbers instead of multiples without a denominator.
