Executive summary
Headline findings
Top 5 moves
AI usage baseline
What is actually happening: corpus size, trend over time, tool mix, where the usage lives, and how confident this review can be in what it saw.
Tool mix
Which AI tools the corpus draws from.
Shadow-AI surface
Which accounts and tiers the usage actually lives on, and what that means for data control.
| Tool | Account tier | Users on this tier | Conversations | Data control implication |
|---|
Coverage honesty box
Included in this review
Not included in this review
Workflow inventory
Every significant recurring workflow found in the corpus, ranked by priority score (frequency × pain/complexity × fit). This is the map the verdict board is built on.
| Rank | Workflow / theme | Conversation share | What the team is really doing | Current workaround | Priority score | Verdict card |
|---|
Verdict board
One card per significant workflow. Every workflow in the inventory above has a verdict here, or an explicit LEAVE. Filter by verdict, effort, or impact to focus the conversation.
| ID | Workflow | Verdict | Recommended action | Effort | Impact | Owner | Tag |
|---|
Risk and exposure register
Sensitive-data-in-chat findings by category only, plus policy gaps. No examples, no quotes. Every risk carries a severity and a one-line fix that reappears in the roadmap.
| Category | Type | Severity | What this means | One-line fix | Roadmap ref |
|---|
Roadmap: 30 / 60 / 90 days and 12 months
Sequenced moves drawn only from verdict cards and risks above. Every item cites the card or risk it traces to. Quick wins are flagged for week one so the review pays for itself visibly.