Partnering with Edra: Context for Agents at Scale
Sequoia by Luciana Lixandru ● Covered by 2 sources
Edra is turning a company’s own tickets, emails and logs into live AI context. That could save teams from re-teaching every bot the same messy business rules.
Based on reporting by Sequoia, by Luciana Lixandru — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Most enterprise AI still starts dumb. A general-purpose model lands in a company full of exceptions, workarounds and tribal knowledge, then needs humans to explain everything it can’t see. That hand-holding is slow, expensive, and it has to happen again when the process changes. Edra’s pitch is to stop doing that work by hand.
The company, backed by Sequoia, is built by Eugen Alpeza and Yannis Karamanlakis, who know the problem from inside Palantir. Eugen spent seven years there and helped build its U.S. commercial go-to-market motion, including the work with AT&T. In 2023, he led Palantir’s AI Platform launch under Alex Karp. Yannis became Palantir’s first Forward Deployed AI Engineer, and the two created that role together to push AI systems from demos into production at scale.
Edra’s idea is simple and a little obvious in hindsight: use the data companies already produce. Support tickets, emails, logs and chat histories become a living knowledge base that reflects how work actually gets done, not how a process document says it should work. The system learns as people use it, and it stays transparent and editable instead of hiding everything inside a fine-tuned black box.
The early focus is on IT service management and customer technical support, where the data is rich and the pain is immediate. Sequoia says the first customers are enthusiastic and expanding fast. That is usually the point where a startup either becomes infrastructure or gets politely admired and ignored. Edra’s bet is that enterprise knowledge is not a document problem. It is a context problem.
Yannis also led a recruiting search engine project that raised placement rates for a staffing firm by 129%. That helps explain why this team keeps coming back to the same theme: AI is useful when it knows the messy reality of the business, not the fantasy version written up after the fact.
My take — AI-written commentary, not fact-checked reporting
This is the rare enterprise AI pitch that doesn’t smell like a demo dressed as a strategy. The black-box crowd loves to pretend models will magically absorb company chaos; Edra is saying, fine, let’s use the chaos as the source of truth. That’s a much more believable business, which is exactly why it will make the hype merchants uncomfortable.
Read more about this at: Sequoia