# Gap Analysis — IT / Helpdesk

These 5 questions CANNOT be answered from your Jira export, Confluence extract, and Okta export alone. They require access to systems without MCP connectors.

1. **Which access requests are pending approval, and are there any accounts that should have been deprovisioned?**
   This requires: Okta admin console (direct) — no MCP connector available.
   Today you would: log into Okta admin console directly and look it up manually.

2. **Which new joiners started this week, and are their provisioning checklists complete?**
   This requires: HiBob (employee data) — no MCP connector available.
   Today you would: log into HiBob directly and look it up manually.

3. **Which cloud access roles have been provisioned, and are any stale or unused?**
   This requires: AWS IAM (cloud access) — no MCP connector available.
   Today you would: log into AWS IAM directly and look it up manually.

4. **Which repository permission changes are pending, and do any need revocation?**
   This requires: GitHub (repo permissions) — no MCP connector available.
   Today you would: log into GitHub directly and look it up manually.

5. **What would a complete view look like?**
   Combining data from Okta admin console, HiBob, AWS IAM, GitHub, plus your existing 3 sources, would give you a 360° view. Which of these gaps causes the most pain in your daily work?

---

**Reflection:** The value of MCP is that it eliminates manual lookups for connected systems. The gap systems listed above represent the manual work that remains. Which one would save you the most time if connected?