Prep each weekly forecast one-on-one with a snapshot diff and question library
Before each forecast one-on-one, this assistant reviews one rep's pipeline, notes what changed since last week and suggests sharp questions. Managers walk in prepared for every rep, not just the loud ones.
Paste this into ChatGPT, Claude or your agent tool. See the prompt
- Deal Review / Forecasting
- intermediate
- Sales leader
- Category
- Deal Review / Forecasting
- Difficulty
- Intermediate
- Est. cost
- Not stated by source
Problem
Managers lack time to walk each rep's pipeline before forecast one-on-ones, so prep skews toward the loud reps and the questions turn generic.
Setup
A Claude Skill run weekly before one-on-ones pulls one rep's open pipeline, diffs commit, best case and upside against last week's snapshot, deep-dives the top commits' recent activity, and picks questions from a library indexed by movement pattern. It outputs a one-page Markdown briefing and never produces a forecast number, auto-shares, or compares reps.
Prompt
Starter prompt by The GTM Agent Weekly, written from the setup described on this page. It is not the source's own prompt; see the source link below for their write-up.
# ROLE
You prepare a sales manager for a weekly forecast one-on-one with one rep. You produce a
one-page briefing with questions. You never produce a forecast number.
# INPUTS
- {{REP}}: the rep's name.
- {{PIPELINE_NOW}}: the rep's open opportunities today (stage, amount, close date, forecast category).
- {{PIPELINE_LAST_WEEK}}: last week's snapshot of the same.
- {{ACTIVITY}}: recent emails, meetings and call notes on the top commit deals.
- {{QUESTION_LIBRARY}}: questions grouped by movement pattern (e.g. pushed date, new to
commit, amount down, no activity).
# TASK
1. Diff this week vs. last week: what entered or left commit, best case and upside; amount
and close-date changes.
2. For the top {{N}} commit deals, summarize the last two weeks of activity in one line each.
3. Match each notable movement to its pattern and pick one or two questions from the library.
# OUTPUT (Markdown, one page)
## {{REP}}: week of {{DATE}}
### What moved
### Top commits: recent activity
### Questions to ask
# GUARDRAILS
- No forecast number, no judgment of the rep, no comparison with other reps.
- Do not share the briefing automatically; it is for the manager.
- Missing snapshot: say so and show today's pipeline only. Replace every {{PLACEHOLDER}} before running. Works with any capable LLM and runner.
Tools
Steps
- Save a pipeline snapshot on a fixed weekly schedule with stable columns.
- Check the requester manages the rep before loading any data.
- Compute week-over-week category movement first, then pull activity for the deals that moved.
- Select questions from a library indexed by movement pattern; say so when none fits.
- Produce a one-page briefing for the manager to read before the call.
Results
The author describes expected time savings for prep; no measured outcome data.
Used in recipes
What goes wrong
Gotchas
- A missed weekly snapshot makes every deal look like it moved; the skill refuses when the gap is too long.
- Schema drift between snapshots produces hallucinated diffs; the skill refuses on column changes.
- Week-over-week movement for ramping reps is mostly noise and can trigger unfair flags.
- Wrong-manager briefings would leak deal context, so the manager check is a hard refusal.
Get 5 like this every week, free.
Real agent setups with sources and failure modes, in one short weekly email.
Free. One email a week. Unsubscribe anytime.
Source and credit
Marius Bughiu, ooligo · newsletter
Source dated 2026-10-11 · last reviewed 2026-10-11. Summarized in our own words; see the source for details.