Israeli startup Encore AI, formerly known as Insait, closed a $30 million Series A led by Team8 and Plaven, according to Calcalist. Garage, Lukatz, and several global financial institutions that are also Encore AI customers participated in the round.
Behavioral Cloning for Customer-Facing Roles
Encore AI’s platform analyzes millions of customer interactions, including phone calls, chats, emails, landing pages, and CRM data, to extract the behavioral patterns and work habits that produce measurable business outcomes. The company then uses those patterns in two ways: building fully autonomous AI agents that handle customer interactions independently, or powering a real-time “Wingman” assistant that coaches human representatives during live calls, according to Calcalist.
The approach differs from most enterprise AI deployments. Rather than training agents on product documentation or scripted responses, Encore AI studies what top performers actually do and replicates those patterns. The distinction matters: most customer-service automation fails because it follows scripts rather than learning from human excellence.
Customers Became Investors
Bank Leumi, Harel Insurance, and Garage are among Encore AI’s customers. Several investors in the Series A round were existing customers who converted to equity holders after seeing direct results from the platform.
“Many investors in this round are also customers, which reflects the value they’re already getting from the platform,” CEO Dvir Ginzburg told Calcalist. “At no point did we feel AI was hurting our business. On the contrary, it accelerated demand.”
The customer-to-investor conversion pattern has appeared in multiple enterprise agent funding rounds this month, a signal that companies deploying agents internally are confident enough to take equity positions alongside their contracts.
Purpose-Built Agents, Not General Copilots
Encore AI’s focus on behavioral cloning for customer-facing roles places it in a growing category of startups building autonomous agents for specific job functions rather than general-purpose AI assistants. The distinction is practical: a sales agent trained on patterns from millions of real interactions starts with an understanding of what works in that role, while a general-purpose model starts with zero domain-specific behavioral context.
The $30M round gives Encore AI runway to expand beyond its current financial services and insurance customer base into new verticals where high-volume customer interactions generate enough behavioral data to train effective agents.