--- name: wealthbox-pipeline-review description: Walk a Wealthbox opportunity pipeline stage by stage to find stalled, unowned or stale deals, separate real risk from clutter, and propose specific updates or follow-ups for approval before any stage or record is changed. Use when the user asks to review the pipeline, find stuck or stale deals, check what needs attention in their opportunities, or prepare a deal review for a meeting. --- # Review the opportunity pipeline A pipeline review answers one question: **which deals need a decision or an action this week, and which are only looking alive?** Stage totals are the easy part. The value is in spotting deals whose stage no longer matches their activity, and recommending the next move. Do not change anything while reviewing; the output is a list of proposals. ## 1. Fix the scope before pulling data Confirm three things with the user, because a wrong scope produces a confident but meaningless review: - **Which pipeline.** Use `GetOpportunityPipeline` to list pipelines, then ask which one if the firm has several. Reviewing two pipelines as one hides stage differences. - **Whose deals.** Only the user's own opportunities, a team's, or the whole firm's. Say which one you used in the output. - **Which window.** Open deals only is the default. Ask whether closed-won and closed-lost should be included for a retrospective. If the signed-in account cannot see the deals the user expects, say so plainly. Do not assume the pipeline is empty or complete. ## 2. Read the stage structure before the deals Call `GetOpportunityStage` for the chosen pipeline. Note the stage order and which stages are terminal. Every later judgement depends on the stage order: a deal sitting in "Proposal" for 90 days means something different from one in "Discovery". If stage names are ambiguous, quote them back to the user rather than guessing their meaning. ## 3. Pull each open opportunity and read its signals For each open opportunity, call `GetOpportunity` and check these signals in order: 1. **Age in stage.** How long since the stage last changed. Use the stage-change date if it exists; if the record only shows created or updated dates, say that the age is an approximation. 2. **Owner.** Is there an assigned owner, and is that person still active on the team? Unowned deals are the most common silent loss. 3. **Next step.** Is there a dated next action or open task. A deal with no next step and no recent activity is stalled regardless of its stage. 4. **Recent activity.** Look for recent notes, meetings or comments via `GetNote`, `GetMeeting` or `GetEvent` on the related contact. A stage that says "Negotiation" with no contact in two months is a status that has drifted. 5. **Amount and close date.** Is there an amount, and is the expected close date in the past? A past close date on an open deal is a near-certain hygiene problem. 6. **Linked contact.** Does the opportunity connect to a real contact record. Orphaned deals are often duplicates. Work through the deals in value order where amounts exist, so the most important ones are checked even if you run short of time. Say if you did not finish the list. ## 4. Classify, don't just list Assign each deal one primary label: | Label | Test | Typical action | |---|---|---| | **Stalled** | No activity or next step, and age in stage well above the rest of that stage | Ask owner for a re-engagement or a close/lost decision | | **Stale date** | Expected close date already passed | Re-date or close, after owner confirms | | **Unowned** | No owner or owner inactive | Assign owner | | **Hygiene** | Missing amount, orphaned, or duplicate-looking | Fix record | | **Healthy** | Recent activity, dated next step, plausible date | No action | Judge "well above the rest of that stage" by comparing deals in the same stage, not against a fixed number. A common working rule is that a deal older than roughly twice the median time in its stage deserves a look, but treat that as a prompt to check, not a verdict. ## 5. Common patterns and traps - **Pipeline inflation.** Many deals sit in late stages with no activity. Usually these were never closed out. Recommend a clean-up pass rather than forecasting them. - **Stage drift.** A deal moved forward on a verbal promise that no record reflects. Ask the owner to confirm the stage before trusting it in the totals. - **Duplicate deals.** Two opportunities on one contact with similar amounts are often a re-created deal. Flag both and ask which is current; never merge or delete. - **Owner is the bottleneck.** One person owns most stalled deals. Note it as a pattern for the manager, not as individual blame. - **Activity without progress.** Lots of notes but no stage movement or next step means busy work. Say so neutrally. - **Missing records is not missing activity.** If notes were captured in another tool or on paper, absence in Wealthbox does not prove nothing happened. Ask before calling a deal stalled on those grounds alone. - **Small amounts crowding out large.** Sort by amount so a few large deals are not buried under many small ones. ## 6. Propose, then wait For each flagged deal, draft the proposed change and its reason: the stage move, the new close date, the owner to assign, or the follow-up task to create. Present them together so the user can approve in one pass. Prioritise a short list of the three or four decisions that matter most this week. ## Output ``` Pipeline: | Scope: | Window: Open deals reviewed: of (value total where recorded) Needs decision this week 1. — , ,