Discovery to Follow-Up Chain: SPIN Prep → MEDDPICC → Buyer Recap

What this chain helps you accomplish

Prepare for a meaningful B2B discovery call, turn the actual conversation into an evidence-backed qualification assessment, and write a follow-up the customer can recognize as accurate.

Four connected AI prompts produce a SPIN Discovery Call Plan, a Verified Discovery Evidence Brief, a MEDDPICC Qualification Scorecard, and a Buyer-Accurate Follow-Up & Next-Step Pack. Each output supplies approved context to the next prompt, with a natural pause between call preparation and post-call analysis.

The chain helps you ask focused, nonleading questions; identify the customer's stated problems and outcomes; distinguish qualification evidence from missing information; and propose a sensible next step without pretending a meeting, budget, or decision has been agreed.

The prompts run in your own AI tool. Promptifi supplies the instructions, not an automatic meeting recorder, transcript processor, CRM integration, or email sender. You review the drafts and separately authorize any customer communication.

When to use it

• Before first discovery: An AE or SDR has a scheduled or planned discovery discussion and wants tailored SPIN questions that help reveal outcomes, not a generic interrogation.
• Immediately after discovery: You have consentfully obtained call notes or a transcript and need a reliable summary of what was actually stated, agreed, challenged, and left unknown.
• During early opportunity qualification: You want an honest MEDDPICC coverage check that separates verified buying process information from seller assumptions.
• Before sending the meeting follow-up: You need a concise recap in the buyer's language, a legitimate proposed next action, and an internal list of gaps to validate later.
• For a multi-attendee or complex B2B sale: You want to preserve who said what and avoid attributing one person's view to the whole buying group.

Not appropriate when: You have no actual call record yet and want the AI to 'score' a fictional conversation, or you are trying to manufacture a buyer commitment, urgency, or forecast confidence. In those cases, use the pre-call prompt only or supply reliable notes after the conversation.

SPIN helps structure discovery questions; MEDDPICC helps evaluate evidence about a potential sale. Neither method, by itself, guarantees that an opportunity is qualified or will close.

Before you begin

Minimum tool: A text-capable conversational AI tool such as ChatGPT, Claude, Gemini, or another organization-approved model. No premium feature, browsing, CRM connector, calendar access, recording integration, or automation is required. Text notes are enough if they are accurate and attributable.

Two stages of inputs: Prepare the pre-call context first. After the actual call, add a transcript or notes when running Prompt 2. Do not invent post-call content in advance. The placeholders below are text to replace, not live CRM or meeting connections.

• {{ACCOUNT_NAME}} — Required. The customer or prospect company, from CRM or your invite. Example: Northline Industrial Distribution. If unknown, identify the meeting account first.
• {{PROSPECT_AND_ROLE}} — Required as a field; can be a verified role without a name. Attendee name, title, and company confirmed from invite or account record. Example: Riley Chen, VP Customer Operations; Avery Patel, Distribution Manager. If uncertain, enter ROLE UNVERIFIED. Never invent an attendee.
• {{OUR_SOLUTION_AND_LIMITS}} — Required. What you sell, relevant capabilities, limitations, and approved claims, from your approved materials. Example: a service that centralizes incoming order inquiries and status visibility; integrations require assessment. If unclear, clarify the offer before the call.
• {{MEETING_PURPOSE_AND_STAGE}} — Required. Why the meeting is happening and the specific question or outcome you aim to learn. Example: first 30-minute discovery to determine whether cross-team handoffs are worth investigating. If no opportunity exists, say 'early discovery'; do not assume a purchase evaluation.
• {{KNOWN_ACCOUNT_FACTS}} — Optional. Dated, attributable research or seller history labeled P1, P2, etc., with source/date and limits. Example: P1 — approved seller notes, Oct. 2026 — distribution expanded to another region. If none, enter NONE; do not claim recent company activity.
• {{PROBLEM_HYPOTHESES}} — Optional. What might be difficult for the buyer, clearly labeled unverified. Example: handoff ownership may be unclear across facilities; not buyer-confirmed. Enter NONE if unknown.
• {{CALL_DURATION_AND_FORMAT}} — Optional. Expected duration, format, and agenda limitations from the invite. Example: 30-minute video meeting; two buyer attendees. If unknown, enter UNKNOWN; a concise adaptive plan is still possible.
• {{EXISTING_RELATIONSHIP_AND_PRIOR_COMMITMENTS}} — Optional. Verifiable interaction history and previously agreed actions from CRM or seller notes. Example: introduced by colleague; no commitments yet. Enter NONE or UNKNOWN otherwise.
• {{CLAIM_AND_DATA_GUARDRAILS}} — Optional. Claims, examples, topics, privacy or recording restrictions, and company policies. Example: do not mention unapproved ROI; customer notes are confidential. If absent, enter NONE PROVIDED and use conservative wording.
• {{FOLLOWUP_SENDER_AND_CONTEXT}} — Optional until Prompt 4. Sender name/title, known recipients, approved product references, and actual reply-to context. Example: account executive following up directly to Riley and Avery. If unknown, mark DRAFT ONLY until verified.
• {{CALL_NOTES_OR_TRANSCRIPT}} — Required only after the call for Prompt 2. Paste a permissioned, accurate transcript or notes with speaker names and timestamps where available. Example: T1 — Riley, call Nov. 5, 2026, 09:12 — 'We reconcile request status manually every Friday.' If a note is a paraphrase, identify it as PARAPHRASE. If no call occurred or notes are unavailable, enter NONE and stop at Prompt 1.

Source labeling: P# = pre-call source or prior note. T# = direct post-call transcript segment or attributable meeting note. Preserve speaker, date, and timestamp if genuinely available. Label DIRECT QUOTE, PARAPHRASE, SELLER OBSERVATION, ASSUMPTION, and UNKNOWN distinctly.

Privacy: Use a company-approved AI environment. Verify consent and legal/organizational rules for recording, transcription, and processing meeting data. Redact sensitive details and only share material you are authorized to provide. If a transcript was generated automatically, review it for speaker attribution and transcription errors.

Evidence standard: Customer-reported statements show what a person said, not independent proof of metrics, budget authority, decision status, or organizational consensus. Lack of information is UNKNOWN, not a negative finding.

Timing: The required effort depends on meeting length, evidence quality, call duration, and the seller's review; no specific time savings have been independently benchmarked.

How to run the chain

1. Start one new AI conversation. Fill and paste the Starting Discovery Context block under Chain Steps, including P# source excerpts and known guardrails. Leave the post-call input as NOT YET AVAILABLE.
2. Copy and run Prompt 1. Review the SPIN questions for relevance, neutrality, and sensible timing. Choose the highest-priority questions rather than reading all fifteen mechanically. Approve the latest SPIN Discovery Call Plan.
3. Conduct the actual customer meeting separately, using your natural conversational judgment and your organization's recording/notes policies. Promptifi and the AI chain do not attend or record the meeting.
4. After the meeting, paste the authorized notes/transcript in the same AI conversation, labeled Call Evidence Input, with T# segments and speaker/date where possible. Then run Prompt 2. Verify quotes and any stated next steps; approve the Verified Discovery Evidence Brief.
5. Run Prompt 3. It must use the latest approved evidence brief and SPIN plan. Review every MEDDPICC rating and its quoted evidence. Approve the latest MEDDPICC Qualification Scorecard.
6. Run Prompt 4. Review the buyer-facing email against the original notes, source IDs, recipient list, and actual commitments. Approve the corrected draft internally. Save the follow-up email and internal seller action list separately.

Approval syntax: APPROVED — [Artifact Name] — v1. Use this approved version in later prompts. For every revised artifact, increase its version number, review the full replacement, and explicitly approve it before proceeding. Seller approval is not buyer approval.

One chat is preferred. If context is lost or you start another conversation, paste the original context, all required complete approved artifacts with versions, and the actual call evidence. Never assume a separate chat can remember the meeting.

If notes are missing: Stop after pre-call preparation or collect a factual retrospective note from the seller; do not create a fictitious transcript or qualification results. If the buyer made no commitment, the follow-up should propose a next step rather than state it as agreed.

Success means: A practical pre-call plan, an auditable post-call evidence record, a qualification map without invented signals, and a useful follow-up email draft. No customer email, CRM update, calendar invite, or automated action is sent by this chain.