Madison Hernandez

Proven team, but CRM has a long history of well-funded AI-native challengers hitting the Salesforce/HubSpot wall. The real moat question: does accumulated conversation context create meaningful switching costs, or can a customer export and move on in a day?

Building on a vectorized database from scratch instead of bolting AI onto Bullhorn was the right hard decision. Most recruiting AI startups try to add a layer on top of legacy data models and it never really works. The semantic matching for non-linear career paths alone is a huge unlock for recruiters. Congrats on the YC batch and the Base10 raise.

Encoded Memory is the differentiator worth watching here. Most agent tools treat every session as day one. If this actually compounds task-level knowledge into reusable programs, you solve the biggest drop-off problem in automation: users building workflows that break the moment something changes. The cloud sandbox is a smart trust architecture choice too.

The open-source angle matters more than people realize for calendar agents. When you give an AI write access to your schedule, being able to audit exactly what it does is a real trust unlock. For founders juggling multiple investor meetings per week, the text-based booking removes genuine friction — the back-and-forth that kills deal momentum is one of those invisible time sinks nobody talks...