YOUR BRAND · YOUR CUSTOMERS · YOUR AI BUSINESS
Sell AI employees under your own brand.
AI Agent News Today
Thursday, November 19, 2026Microsoft Foundry: built-in governance for agent fleets
What changed: Microsoft published a Foundry product post describing built-in governance for enterprise AI agents — agent identities via Microsoft Entra, model and agent deployment policies, tool and network controls, cost visibility and token/quotas, and safety guardrails — as generally available or preview features for customers to start using.
Why it matters: Platform teams and operators can apply the same identity, access, cost and safety policies across many teams instead of stitching controls together after the fact, which lowers the compliance and operational work needed to move agents from prototype to production.
Try/watch: Start with a single subscription: assign agent identities, enable model/agent deployment policy checks, and configure a token or spend threshold so you can see how Foundry surfaces policy violations and cost spikes before broader rollout.
Azure Monitor: unified agent observability and end-to-end tracing (preview)
What changed: Microsoft announced a preview of unified agent observability in Azure Monitor Application Insights that uses OpenTelemetry trace context and the OpenTelemetry GenAI semantic conventions to trace requests across multiple agents, platforms, models and tools. The post explains supported providers, trace correlation, and invites teams to enable telemetry to follow a request end-to-end.
Why it matters: If your systems use multiple agents or delegate work between agents, you no longer have to stitch disparate logs to answer where a failure or cost spike happened — traces can show which agent, tool or external call slowed or failed, making debugging, cost analysis, and incident response practical.
Try/watch: Enable telemetry on a small cross-agent workflow (for example: an order-risk check that calls two agents) and verify traces propagate across handoffs; redact PII in telemetry and validate trace fidelity before increasing sampling or retention.
Oracle Life Sciences AI Agents: agentic coordination for clinical trials
What changed: Oracle announced Life Sciences AI Agents to help clinical operations monitor portfolio health, surface operational risks, and recommend next actions for review; the release positions the capability for early adopters and emphasizes human review and approvals for any agent recommendations.
Why it matters: Clinical teams and sponsors can pilot agents that aggregate signals from multiple trial systems and suggest operational actions (site selection, enrollment risks, protocol deviations) while keeping humans in the approval loop — a practical path for regulated workflows where auditability and human sign-off matter.
Try/watch: Run a narrow pilot that uses agents only to surface-ranked issues (not to auto-execute changes), map every suggested action to an approval owner, and capture the agent’s reasoning and audit trail to evaluate regulator and compliance readiness.
Stitch AI: agentic assistants inside a banking operating system
What changed: Stitch released Stitch AI (PR release) to embed Copilot-style assistants, ready-made operational agents (onboarding, credit decisioning, collections) and an Agent Harness directly inside its banking operating system, with scoped permissions, maker/checker approvals, audit trails and PII masking before model access.
Why it matters: Banks that must keep strict controls and audit logs can let agents act closer to the ledger because the agents operate through the core system (not on copies of records) and record approvals and reasoning — this reduces integration risk when regulators and risk teams require provable controls.
Try/watch: Pilot a non-customer-impact workflow (for example: an onboarding checklist where the agent prepares recommended steps but requires human sign-off), validate masking and audit logs, and test switching model endpoints to confirm the platform keeps the same controls regardless of model choice. Microsoft Foundry: Govern your AI fleet with confidence on Microsoft Foundry. Unified agent observability with end-to-end tracing in Azure Monitor. Oracle Life Sciences AI Agents Bring AI-Assisted Operational Intelligence and Coordination to Clinical Trials. Stitch Launches Stitch AI, Bringing Agentic AI Directly Into Banking Operations.
YOUR BRAND · YOUR CUSTOMERS · YOUR AI BUSINESS
Sell AI employees under your own brand.
Give businesses AI teams that own ongoing roles. Set your pricing, delegate routine coordination and build a recurring service.