Know which dealsare real.
Four Claude Skills read your open pipeline, score whether each deal is truly committed, and draft your deal-review brief and rescue nudges. It draws the line between the floor you can bank and the deals propping up an inflated number — and won't count a deal with no next step or a dead close date as committed, however large. It reads and drafts; you update the CRM and send.
Most forecasts are propped up by deals that aren't real.
Your forecast counts deals that stalled weeks ago. Leadership plans — and you commit — on a figure held up by deals that were never closing this period.
A deal with no scheduled next action, a past-due close date, or weeks of silence is already slipping. But it sits in the pipeline at full value until someone finally admits it.
Every forecast call is the same scramble: re-reading deals one by one, guessing which are real, with no honest floor and no clear answer to which deal to work first.
Score a deal — and watch a slipping one inflate the forecast.
This is the scoring core, live. Tune the $120k deal and watch its verdict, your committed floor, and the CLEAN/INFLATED call move with it. The commit gate is the honest part: drop its next step, activity, or close date and it slips — however high the score climbs everywhere else.
Pipeline health · live demo
Tune the $120k deal at the top. Drop a blocker and watch it slip — and the forecast inflate.
Committed floor
$105,000
of $360,000 pipeline
Forecast: INFLATED
$145,000 slipping (40% of pipeline)
Work this first: Atlas Mfg ($120,000, SLIPPING) — the most forecast dollars you can still save.
Atlas Mfg · $120,000
SLIPPING · score 38
The commit gate is the point: a deal that fails next step, recent activity, or close-date realism is SLIPPING however large — it can’t prop up the forecast. Same math as the engine and workbook. It reads and drafts only — you update the CRM and send.
One pipeline read, four Skills, from inflated to honest.
The scout reads your open deals via your HubSpot or GoHighLevel MCP and scores each, returning COMMIT / BEST-CASE / AT RISK / SLIPPING, your committed-dollar floor, the CLEAN or INFLATED roll-up, and the deal to work first.
The rescue drafter writes the move that makes a deal real again — a next-step proposal, a re-engagement, a stakeholder reach — in your voice, grounded in the actual deal, with anything unconfirmed flagged [confirm].
The architect turns the read into the deal-review brief — the floor, what's slipping and why, the work-first list, talking points — and can stage it as an unshared doc in your own Drive. It never edits the CRM.
Between reviews the tracker re-scores the pipeline and flags what newly slipped, what firmed up into COMMIT, and where a CRM stage no longer matches reality — so the floor stays real instead of drifting.
Plus a runnable Python scoring engine with a worked sample, a spreadsheet that reproduces the same math and commit gate, the deal-review brief template, and two playbooks — the end-to-end pipeline-hygiene workflow and a CRM MCP setup & safe-deal-review guide.
Three rules the floor holds to.
Scoring runs on your actual open deals through the MCP. It never fabricates a deal, an amount, a close date, or activity — and stops rather than guessing if the CRM connection is missing.
It reads and drafts only. It may stage the review brief as an unshared doc in your own Drive, but it never edits a deal, moves a stage, changes an amount or close date, sends a nudge, or deletes. Those are yours.
A deal with no next step, a past-due close date, or weeks of silence is SLIPPING — and can't count as committed forecast, however large. Refusing to bank a number on deals that aren't real is the point.
An orchestration layer on top of your CRM — honest about its limits.
- A repeatable read that scores every open deal and gives you the committed floor you can actually bank.
- The orchestration layer — Skills, scoring, a template, and playbooks — on top of your CRM (HubSpot or GoHighLevel) + Claude.
- Built to keep the forecast honest: it won't count a deal that fails a reality check as committed, however large.
- Not a CRM — you bring your own HubSpot or GoHighLevel account and MCP connection.
- Not an auto-pilot: it reads and drafts only, stages the brief unshared, and never edits a deal, moves a stage, changes an amount or close date, sends, or deletes on its own.
- Not a guarantee a deal closes, and not financial or accounting advice — your forecast is your own call.
Built for whoever has to forecast honestly.
- A founder, sales lead, or RevOps owner running a pipeline you have to forecast and defend.
- Tired of a number propped up by stalled deals — and of guessing which deals are real at every review.
- Willing to connect your CRM MCP and stay the one who updates deals and sends the follow-ups.
- Want a bot that auto-updates stages, edits close dates, and sends follow-ups on its own.
- Won't connect a CRM, or review drafts before sending.
- Need a full revenue-operations platform across teams and quotas — this scores deal health, not org-wide RevOps.
Win it, start it, sound like you.
Reclaim the deals that already slipped away — re-engage lapsed and closed-lost in the same CRM.
Start the deals you win cleanly — score kickoff-readiness and draft the whole onboarding packet.
Codify your voice so the rescue nudges and the brief sound like you — not generic AI.
The honest answers, before you buy.
A CRM — HubSpot or GoHighLevel (it works with either) — for your open pipeline. Optionally Google Calendar (to confirm a deal's next-step meeting is really booked) and Google Drive (to stage the deal-review brief as an unshared doc). You bring your own accounts and connect them to Claude via MCP; the Forecast Floor is the orchestration layer on top — it never includes or resells your CRM.
It scores each open deal on six signals — next step defined, recent activity, champion engaged, qualification confirmed, close-date realism, and whether the stage matches reality — each weighted, summing to 100, for a 0–100 score and one verdict: COMMIT (≥ 78, gates pass), BEST-CASE (55–77), AT RISK (< 55), or SLIPPING. The honest part is the commit gate: three of those signals are commit-blockers (next step, recent activity, close-date realism), and if any is at or below 1, the deal is SLIPPING no matter how high the score — a deal that fails a reality check can't count as committed, however large. In the worked example a $120k deal with no next step, gone silent, and a past-due close date scores 38 and reads SLIPPING.
No — and that's deliberate. It reads your pipeline and drafts next-step nudges, a deal-review brief, and the forecast roll-up, and it may stage the brief as an unshared doc in your own Drive. It never edits a deal, moves a stage, changes an amount or close date, sends a message, or deletes anything. Updating the CRM and sending are yours to do after you review.
The Sales Enablement Battlecard Builder and the Synthetic Customer Research Panel help you win individual deals; the Investor-Ready Metrics Pack reports company metrics to investors. The Forecast Floor sits between them: it's a connected workflow that reads your live pipeline and grades it deal-by-deal so your forecast is honest — which deals you can bank, which are slipping, and which one to work first. It pairs naturally with the CRM Win-Back System (lapsed deals) and the Client-Onboarding Runway (the deals you win), all on the same CRM.
Four Claude Skills (pipeline-health scout, deal-rescue drafter, deal-review-brief architect, forecast-hygiene tracker), a runnable Python scoring engine with the worked sample, the same scorer as an Excel workbook, a deal-review brief template, and two playbooks — the pipeline-hygiene workflow and a CRM MCP setup-and-safe-deal-review guide. You supply the CRM account.
No on both. It tells you which deals are real, what's slipping, and what to work — but whether a deal closes depends on you and the buyer, and outcomes are never guaranteed. And the COMMIT/SLIPPING verdict is forecast hygiene, not financial or accounting advice: your forecast and how you report it are your own call.
Stop forecasting fiction.
Know your floor.
One purchase, lifetime access, 12 months of updates. $149, once. Bring your own CRM MCP.
Sold by RedHub AI LLC · Secured by Stripe · redhub.ai