Encore AI raises $30M to build AI agents that learn from customer calls
- ID
- 8762
- Status
- summarized
- Published
- 29 Jul 2026, 10:41 PM
- Fetched
- 30 Jul 2026, 12:09 AM
- Provider
- TechCrunch
- Category
- technology
- Original URL
- https://techcrunch.com/2026/07/29/encore-ai-raises-30m-to-build-ai-agents-that-learn-from-customer-calls/
- Source URL
- https://techcrunch.com/feed/
Summary
- Score
- 5.5
- Created
- 31 Jul 2026, 4:33 PM
- Tags
- Audience
- ai_agent_userssaas_foundersai_ml_learners
What happened
Encore AI (formerly Insait IO), founded in 2022 by CEO Dvir Ginzburg, raised a $30M Series A led by Team8 to build AI voice/text agents that learn from companies' recorded customer calls, emails, and CRM data. The platform performs 'interaction mining'—dividing conversations into stages, identifying which approaches drove successful outcomes, and training agents that replicate the most effective employee playbooks, including specific jokes and anecdotes.
Why it matters
If you run a sales or support team with call recordings and CRM data, this is a concrete example of how conversation analytics can be turned into autonomous AI agents that mirror your best performers' tactics. For builders, the 'interaction mining' approach—segmenting conversations by stage and correlating with outcomes—is a design pattern worth considering for any voice/text agent project, not just a vendor pitch.
Discussion angle
The 'interaction mining' concept is the interesting part: rather than training agents on generic data, they segment real customer conversations by stage and tie outcomes to specific conversational moves. Is this approach replicable with open-source tools, or is the moat purely in having access to enterprise call data?