Encore AI secures $30 million to train voice agents on customer conversations

Encore AI has raised $30 million in a Series A led by Team8 to expand AI agents trained on companies’ own customer interactions. Founded in 2022 as Insait IO, the startup now serves more than 40 enterprises worldwide, most of them financial institutions.

Why interaction history matters

Generic voice agents may know a product but still miss the tactics that move a particular customer relationship forward. Encore analyzes how employees handle calls, emails and messages, then identifies approaches associated with successful sales or support outcomes.

The system can expose friction, weak handoffs and ineffective stages in existing workflows. Its agents can communicate with customers by voice or text, operate autonomously, or assist employees by recommending responses and tactics during live conversations.

From conversations to playbooks

CEO Dvir Ginzburg calls the method “interaction mining.” The platform collects call recordings, emails and text messages, connects them with CRM records, divides interactions into stages and evaluates which parts advanced or obstructed the process.

“The agent we build is a package of many different playbooks that have worked throughout the process.”

Encore says this can preserve effective details from human interactions, including examples, anecdotes and even jokes used by relationship managers. Annual recurring revenue has grown more than fivefold since its seed round less than 18 months ago, although the company did not disclose revenue or valuation.

Planven, Lukatz, Garage, banks and insurers also joined the round. Some participating financial institutions first adopted the product and later invested. Encore will use the capital to expand US sales and deploy the platform at more large financial institutions.

Salesforce, SAP, Zoho and HubSpot could build comparable functions around their CRM data. Encore argues that its advantage lies not in data access alone, but in making historical conversations the foundation of agent behavior.

For businesses, the practical lesson is to treat customer conversations as operational data before deploying agents. Clean recording policies, CRM integration and measurable outcome labels will determine whether automation reproduces proven expertise or merely adds another generic interface.