Encore AI has raised $30 million in a funding round led by a firm called Team8, with additional backing from Planven, Lukatz, Garage, and several banks and insurance companies. Some of those financial institutions became investors after trying out Encore's software first. TechCrunch

The company was founded in 2022 as Insait IO by CEO Dvir Ginzburg and has since changed its name to Encore AI. Its technology, which Ginzburg calls "interaction mining," listens to and reads recorded phone calls, emails, and text messages between a company's employees and its customers. It then checks those conversations against the company's sales records to figure out which ways of talking to customers tend to lead to good results, like a sale or a resolved problem. The technology is patented. StreetInsider

The platform uses what it learns to power AI voice agents that can work in two ways. They can talk to customers directly by phone or text on their own, or they can assist human employees during live conversations by suggesting what to say next. This setup means Encore is not simply replacing human workers but trying to copy the habits of the best-performing ones and share those habits across the whole company.

Encore says it has more than 40 large business customers worldwide, most of them banks and other financial companies. Its recurring revenue has grown more than five times since its earlier seed funding round, which closed less than 18 months ago. The company plans to use the new money to expand its sales operations in the United States and bring its software to more large financial institutions. TechCrunch

It is worth noting that several of the investing banks and insurers came in only after using the product. In the software world, a customer deciding to also become an investor is not unusual. But it carries more weight in banking and insurance, where strict rules about data protection and record-keeping mean that fewer vendors even get through the door for a trial. A bank using an AI system that processes its call recordings and customer messages has already passed substantial legal and security checks before a single conversation is analyzed. The fact that these institutions then chose to invest their own money suggests the software met not just the requirements to be purchased but the standard of being genuinely useful in day-to-day operations.

Encore's approach also differs from how most conversational AI products are built. The common method starts with a general-purpose AI language model and gives it access to a company's documents and help pages so it can answer questions. Encore's method is more like learning by example. Instead of training on written guides, it trains on the outcomes of real conversations that human employees had, using the sales records to confirm whether each interaction worked. This is harder to build because it requires pulling together and organizing different types of communication across phone calls, emails, and messages. But it produces AI agents whose behavior is based on what has actually worked in practice rather than on a summary of best practices.

The focus on financial services is both an advantage and a limitation. Banks and insurers generate huge volumes of customer interactions, already record and log calls as standard practice, and pay high costs for human agents. Those factors make Encore a natural fit. They also set a high bar. Any AI agent talking directly to customers in a regulated industry must follow rules about disclosing that it is a machine, protect sensitive financial data under laws like the U.S. Gramm-Leach-Bliley Act, and keep a record of every action it takes. Encore's patents and customer base suggest it designed its software with these requirements in mind from the start rather than adding compliance features later.

The $30 million is modest compared to the large sums flowing into AI infrastructure companies, but it is appropriately sized for a business software company entering an industry with long sales cycles and high-value contracts. A fivefold revenue increase in 18 months signals early traction, though the actual dollar amounts are not public. The real test ahead is whether Encore's expansion in the United States can repeat the pattern seen in its earlier stage, where the people putting in money were the same people whose customer data and budgets were at stake.