Sequoia is backing Edra, a startup that converts existing enterprise data, support tickets, emails, logs, chat histories, into a living knowledge base that AI agents can actually use. The problem it solves is concrete: every general-purpose AI deployed inside a company starts from zero, requiring expensive forward-deployed engineers and manual documentation that becomes stale the moment a process changes. Edra skips that cycle entirely by learning from data the company already generates.
The founders are Eugen Alpeza and Yannis Karamanlakis, both former Palantir. Eugen led U.S. commercial go-to-market including the AT&T deployment, one of Palantir's most complex, and launched Palantir's AI Platform under CEO Alex Karp in 2023. Yannis held the first Forward Deployed AI Engineer role at the company and previously shipped a recruiting search engine that raised placement rates 129% for a staffing firm. They have known each other 13 years. The first production use cases are IT service management and customer technical support, two domains where ticket data is dense and automation ROI is immediate.
What makes the full piece worth reading is not the funding announcement. It is the technical framing around why black-box fine-tuning fails at the enterprise level and how Edra's transparent, editable knowledge layer changes the economics of agent deployment. The architecture argument, not the outcome, is the story.
[READ ORIGINAL →]