Conduct Raises $60M to Build the Agentic OS That Enterprises Actually Need

Frequently asked

What did ex-Palantir engineers build before, and why does that background matter for enterprise AI agents?

Palantir built operational data platforms for defense and intelligence clients — environments where data pipelines had to be reliable, auditable, and failure-tolerant under adversarial conditions. That engineering culture, not just the résumé, is what Conduct's founders bring: a preference for deployment contracts over demos, and an understanding that enterprise value is captured in the unglamorous coordination layer, not in model benchmarks.

Why would SAP invest in an agentic OS startup instead of building the coordination layer itself?

SAP's core business is workflow incumbency — it owns the process layer in thousands of large enterprises. Building a general-purpose agentic OS would require competing with its own integration partners and cannibalizing consulting revenue. Investing in Conduct lets SAP embed agentic coordination into its ecosystem without owning the liability of building it. The investment is a distribution bet, not a technology bet.

What is the strongest argument against the agentic OS thesis Conduct is betting on?

The strongest counter is that foundation labs — OpenAI, Anthropic, Google — are building orchestration natively into their APIs and will commoditize the coordination layer before Conduct reaches enterprise scale. If model providers solve agent handoff, state tracking, and tool integration at the API level, a standalone agentic OS becomes redundant middleware. Conduct's answer to this is Palantir-style deployment depth, but that advantage only holds if enterprises choose vendor lock-in to an orchestration layer over native lab tooling.

More on this wire