MEDDPICC deal analyzer for B2B sales qualification
Build a sales-deal analysis dashboard that scores B2B deals against the MEDDPICC qualification framework by reading call transcripts. It's for account executives and sales managers who want a fast, evidence-based read on where a deal stands and what to do next. Trigger it with a single private webpage route (space-authenticated) that serves the dashboard. No cron or email trigger is needed. The dashboard is the centerpiece: a user names a deal, then adds one or more call transcripts for it — supporting both pasting raw text and attaching files (accept common transcript formats like .txt, .vtt, .srt, .md, and .pdf). Uploading more calls later for the same deal should merge into that deal rather than create a duplicate, so every transcript ever added is kept and re-analysed together into one running overview. Let users reopen any saved deal from a folder/sidebar, add further transcripts to an open deal and re-analyse, and remove a deal. At a high level, build these steps: A webpage step that renders the whole dashboard UI (deal folder, upload/paste inputs, and the analysis view). An intake step that receives the deal name and transcripts, extracts text from each file (including PDFs and subtitle formats), accumulates them per deal in persistent storage, and kicks off a background analysis job — responding immediately so the UI can poll for results, since analysis can take a minute or two. An AI analysis step that runs one structured pass across all of a deal's calls together, producing a combined health score, a per-pillar MEDDPICC breakdown with supporting evidence, green/red flags, areas to dig deeper, and suggested questions for the next call. A library step that lists saved deals and returns a single deal's full analysis on demand. A delete step that removes a deal and its stored transcripts. Persist deals and their transcripts in durable storage so they survive across runs and are shared by the workflow. Keep all routes private to the space unless asked otherwise. For the AI analysis, use Anthropic Claude — or your preferred LLM provider — via a connector so no keys live in code. PDF/subtitle text extraction can use any suitable library. Tools used: Anthropic Claude API, unpdf
What this prompt builds
A sales-deal analysis dashboard that scores B2B opportunities against the MEDDPICC qualification framework by ingesting call transcripts. Built for account executives and sales managers who need a fast, evidence-based assessment of deal health and next steps. The workflow aggregates transcripts per deal, produces pillar-by-pillar breakdowns with supporting evidence, and surfaces gaps and suggested questions.
The problem
Account executives and sales managers struggle to consistently qualify complex B2B deals against a rigorous framework like MEDDPICC, often missing critical information buried across multiple sales calls. This workflow reads call transcripts and scores each deal against all eight MEDDPICC pillars — Metrics, Economic buyer, Decision criteria, Decision process, Paper process, Identify pain, Champion, and Competition — surfacing exactly which information is present, which is missing, and what questions to ask next. It aggregates every transcript ever added to a deal into one running analysis, so teams get a fast, evidence-based read on deal health without manually combing through notes.
Solution and impact
Sales teams save hours of manual qualification work and gain a shared, objective view of every deal's strengths and gaps. The workflow delivers a combined health score, pillar-by-pillar breakdowns with quoted evidence, green and red flags, and tailored questions for the next call — so reps can focus on filling gaps and moving deals forward with confidence. By keeping a running analysis that updates as new calls are added, it ensures no critical detail is lost and every stakeholder understands where the deal truly stands.