Partnering with Cymphony: Security Unlocks Adoption
Sequoia Bogomil Balkansky and Dean Meyer ● Covered by 2 sources
Sequoia says Cymphony helps enterprises see what AI agents can access and do. That’s the blocker keeping a lot of agent projects stuck in pilot mode.
Based on reporting by Sequoia, Bogomil Balkansky and Dean Meyer — 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
Enterprise AI has hit a very specific wall. Companies want agents doing real work, but security teams keep asking the same awkward questions: what can this thing reach, and how do we know what it did? Sequoia’s bet on Cymphony starts there, not with model quality or cost, which is a useful tell about where enterprise adoption is actually getting stuck.
The core problem is that agents don’t fit the old security model. They can be set up in minutes, run constantly, and inherit access from people and applications across systems that were never built to agree with one another. That makes them useful. It also makes them hard to approve. If a team can’t see whether an agent, acting for a given employee, can reach a sensitive file right now, the answer from security is going to be no.
Cymphony is trying to map that mess continuously. It tracks both employees and agents across an enterprise and builds a picture of who and what can reach which systems and information. That is more than just collecting alerts in one place. The harder part is stitching together the many accounts one person may have across disconnected tools, plus the access agents inherit along the way.
The company’s pitch is that security has to move beyond detection and into workflow. So instead of stopping at “here’s the problem,” it also recommends remediation and can automate parts of it, while coordinating with the people who need to act. That seems to be resonating: customers named in the source include Syngenta, KKR, Cass Information Systems, and Athennian. Sequoia says it led Cymphony’s seed round and is now co-leading its Series A.
The founders also fit the sort of profile enterprise security investors love: Shy Dekel, Idan Berkovits, and Edi Gotlieb all came through Israel’s Talpiot program, with backgrounds spanning Unit 8200, the Office of the Prime Minister, Apple, and the Israeli Ministry of Defense. Sequoia’s broader point is simple enough: every major software shift forced security to rebuild around a new unit of work. AI is doing that again, except this time the unit of work can act on its own.
My take — AI-written commentary, not fact-checked reporting
This is the part of AI that actually deserves the budget: boring control layers, not louder demos. The industry keeps selling autonomy like it’s candy, then acts surprised when enterprises want a seatbelt, a dashboard, and an audit trail. Sensible people call that security; startups call it the real product.
Read more about this at: Sequoia