Churn Save Play Drafter

What it takes in

The inputs you provide on each run, validated before anything executes.

2 tasks, run in order.

Each task passes its result to the next. Here's what each does and what it produces.

1. 01
   Diagnose the risk. From {{accountContext}}, identify the top likely churn drivers for this account, citing the specific evidence in the provided context for each (usage trends, sentiment signals, support history, stakeholder changes). Rank the drivers by likely impact. Then assess overall severity given the renewal window: if {{renewalDate}} is provided, factor in how much runway remains before renewal; if {{contractValue}} is provided, note the revenue at stake. Do not invent facts not present in the provided context; if the evidence for a driver is thin, say so. Present this as a readable diagnosis, never JSON.

   Expected output

   A readable risk diagnosis with no JSON and no code blocks containing: Churn Drivers (the top 2 to 4 likely drivers, each with a heading, the specific evidence from the provided context, and a likely-impact rating of High, Medium, or Low); and Severity Assessment (an overall risk rating of High, Medium, or Low with a short rationale that factors in the renewal window and revenue at stake where provided).
2. 02
   Draft the save play. Using the diagnosis from Task 1, propose 2 to 3 intervention options ranked by likelihood of working, each with what it involves, why it could work against the diagnosed drivers, and the effort required. Recommend one play and explain why it best fits the drivers and the renewal window. Then draft an outreach email to the account's champion that acknowledges the situation without sounding alarmed, offers something concrete, and ends with one clear ask. Finally, prepare call talking points for the save conversation, including a scripted response for when the customer says "we're evaluating alternatives". Ground everything in the evidence from {{accountContext}}; do not promise capabilities, discounts, or concessions not implied by the inputs. Present as readable, ready-to-use materials, never JSON.

   Expected output

   A readable save play with no JSON and no code blocks containing: Intervention Options (2 to 3 options ranked by likelihood of working, each with a heading and bullets for what it involves, why it could work, and Effort of High, Medium, or Low); Recommended Play (the chosen option with a short rationale tied to the drivers and the renewal window); Outreach Email (a Subject line plus a ready-to-send body of roughly 100 to 180 words addressed to the champion, ending with one clear ask); and Call Talking Points (bullets for the save conversation, including a scripted response to "we're evaluating alternatives").