Blog
Insights from the Roster team on organizational harness, agentic workflows, and the systems that help agents resolve who should act.
The Organizational Harness: Why Org Context Resolution Is the Critical Path to Agentic Workflow Adoption
A Roster position paper on identity, accountability, and organizational context in agentic AI: why the organizational harness connecting every autonomous action to an accountable owner is the critical path from pilot to production.
Read article →Your AI Agent Approval Workflow Has a Name Hardcoded in It. That's the Bug.
Most AI agent approval workflows still route to a name someone typed in once. Here's what dynamic, directory-based approval routing looks like -- plus service ownership and MCP-callable approval workflows.
Read →AI Agent Governance: Why "Who Should Handle This" Is the Part Nobody Solved
Identity, authorization, audit, and human-in-the-loop are the four pillars of agent governance everyone talks about. None of them answer who should actually be asked — that gap is resolution.
Read →How Roster Fits With Your Platform
A companion piece to our platform survey: what actually changes, category by category, when Roster sits alongside the tools you already run.
Read →How Platforms Handle Participant Resolution and Org Context Today
A formal, category-by-category survey of how agent orchestration frameworks, workflow automation platforms, iPaaS, BPM/ITSM suites, service catalogs, authorization engines, and DIY methods handle "who should act on this" today — plus where Roster fits alongside them.
Read →Human-in-the-Loop Is Not Enough: Agents Need to Know Which Human
Every agent framework now ships a pause-for-a-human primitive. None of them, on their own, answer which human. Why HITL checkpoints and org context resolution are different problems — and why Roster is built to answer the second one.
Read →From Directory Data to Runtime Decisions: Resolving People, Groups, Roles, and Delegates
How Roster actually connects to Entra ID, Okta, Google Workspace, Workday, LDAP, and CSV: what gets read, what gets cached, what never gets written back, and how delegations layer on top without touching the directory.
Read →Why Hard-Coding Approvers Breaks Agentic Workflows
Hard-coding an approver into a workflow works right up until it doesn't — someone goes on leave, changes teams, or leaves the company. Here's what actually breaks, and why the fix isn't a bigger config file.
Read →What Is Participant Resolution?
Every agentic workflow eventually asks who should approve, own, or be notified next. Here's what participant resolution is, how it differs from identity and authorization, and how Roster resolves it in three steps.
Read →The Missing "Who" Layer in AI Agent Workflows
Agent frameworks are good at planning, tool calling, and pausing for human review. Almost none of them know who that human should be. Here's why pausing and resolving are different problems, and what actually closes the gap.
Read →