The Identity Control Surface stops being a diagram

This is the Identity Control Surface problem made product. As multi-agent workflows proliferate, the question of what any given agent is permitted to do, and under whose authority, becomes the governance fault line. AI Passport positions ServiceNow as the runtime enforcement point across agent fleets, including MCP servers that sit outside the primary application boundary. Shadow AI detection closes the registration gap: agents that were never formally enrolled now surface. Enterprises building agentic workflows on ServiceNow gain a concrete mechanism for non-human identity governance, but also take on a new dependency: ServiceNow becomes the authority layer, not the enterprise's own identity stack.

Workday is moving from agent registry to agent access governance, the distinction that separates a catalog from a control plane. C1.ai's technology extends that control surface beyond Workday's own perimeter into the broader enterprise system estate. Combined with the Accenture forward deployed engineering program announced the same week, Workday is simultaneously building the governance layer and accelerating client deployment into it. For enterprises evaluating agentic identity architecture, this consolidation signals that HR and finance system vendors intend to own the Identity Control Surface for workforce agents, not defer it to standalone IAM providers.

Forward deployed engineering is a production pressure mechanism: it moves the deployment bottleneck from the vendor roadmap to the client implementation team. For the Compiled Corporation lens, this matters because HR and finance are where workforce decision-making is encoded, and Workday agents operating at that layer begin to automate the execution of compensation, resource allocation, and compliance workflows. The program also reveals where the Decision Surface currently sits for most enterprises: engineers are still required on-site to bridge the gap between agent capability and production readiness. That gap is the current market.

The Itaú architecture illustrates a pattern emerging among regulated-industry adopters: a public-facing AI interface paired with a dedicated internal governance layer. The separation is structurally significant. ia.i is the Janus Brand surface, the AI persona customers and employees encounter. IARA is the Identity Control Surface, the platform that manages what the AI is permitted to do within a regulated compliance envelope. Enterprises in financial services, healthcare, and other regulated sectors can read this as a reference architecture: brand-facing agent identity and compliance-facing agent governance are separate design problems requiring separate tooling.

This is a governance architecture choice, and it runs counter to the opaque trustee structure surfaced in the concurrent Wall Street Journal reporting on Anthropic's Long-Term Benefit Trust. The external watchdog arrangement creates an independent audit channel with publication rights, giving it teeth that advisory relationships typically lack. For enterprises selecting foundation model providers, the existence of an external auditor with genuine access is a due-diligence signal: it introduces an accountability layer outside the vendor's control. The Janus Brand implication is direct: Anthropic's safety-first positioning now carries an external verification mechanism, which either validates the brand claim or surfaces gaps in it.