Process

The Anatomy of a Broken Approval Workflow

Approval workflows fail in predictable ways. Understanding the failure modes. not just the symptoms, is what separates a durable fix from one that breaks again in six months.

The Lobbi Delivery Team
February 10, 20262 min read

The Lobbi Delivery Team

Operational Systems Engineering

Every organization has an approval workflow that nobody trusts. Requests disappear into inboxes. Approvers claim they never received anything. The same item gets approved twice or not at all. The team works around it, and the workaround becomes the process.

The Four Failure Modes

Broken approval workflows almost always fail in one of four ways. First: no single source of truth. The request exists in an email, a spreadsheet, and a chat message simultaneously, and the versions diverge. Second: no escalation path. When an approver is unavailable, nothing routes. The request stalls until someone chases it manually. Third: no audit trail. When a decision is later questioned, there's no record of who approved what, when, or on what basis. Fourth: no feedback loop. The requester has no visibility into where their request sits. They follow up by sending more messages, which creates more noise.

What a Fixed Workflow Actually Looks Like

A well-designed approval workflow has a canonical state: one record in one system that is the authoritative version. It has defined routing logic that handles delegation and escalation automatically. It notifies in the channel where approvers actually work, Teams, not a separate portal. And it captures every state transition with a timestamp and actor.

The technology is secondary. Power Automate, a custom Azure Function, or a Dynamics workflow can all implement this correctly. What matters is designing the logic before choosing the tool, mapping the states, the transitions, the exception paths, and the audit requirements. Most broken workflows were built by starting with the tool.

Topic clusters