Turn customer calls into a positioning and messaging refresh

Start by choosing 8-12 calls that represent different deal outcomes and lifecycle stages, not only the happiest customers. Include closed-won deals, closed-lost deals, active opportunities, onboarding calls, renewal discussions, and at least one call where the buyer pushed back hard. Export the transcript or call summary from Gong and capture metadata in Airtable: account segment, persona, company size, deal stage, outcome, product area, competitor mentioned, and call date. Mark whether each transcript is approved for internal analysis and whether any customer names or sensitive implementation details need redaction. The goal is to create a research sample that reflects the market conversation, not a highlight reel.

Before asking AI to summarize anything, define what you want extracted from the calls. Create fields in Airtable for buyer words, pain theme, desired outcome, switching trigger, objection, competitor comparison, confusing product language, proof requested, emotional intensity, evidence link, and confidence. Then open Claude, paste the prompt below, and use your sample metadata plus the schema fields as the input so Claude can refine the extraction rules before analysis starts. Copy the output from Claude back into Airtable or Google Docs as the approved extraction schema and redaction rules for the rest of the workflow. QA check: every field should have a clear definition, an example value, and a rule for when the field should be left blank.

Use Gong trackers or search to find moments where buyers discuss pain, urgency, budget, alternatives, implementation risk, proof, and decision criteria. Export the selected excerpts with transcript links and metadata, then open Claude and paste the prompt below with the excerpts attached in batches. Ask Claude to extract exact phrases first, then group them into themes only after the evidence is captured. Copy the output into Airtable with the source call, persona, deal outcome, supporting quote, and confidence score. Do not let the model rewrite rough customer language into polished marketing copy at this stage.

Move the extracted themes into Google Sheets and score them using frequency, deal importance, target persona relevance, urgency, revenue relevance, proof availability, differentiation potential, and whether your product can credibly own the message. Open Claude and paste the prompt below only after the first human scoring pass, using the sheet export as input. Ask Claude to challenge weak scores, identify unsupported themes, and flag themes that look emotionally strong but commercially narrow. Copy the output from Claude back into the sheet as reviewer notes instead of automatically overwriting the scores. The PMM should select the top themes by balancing repeated evidence with strategic relevance.

Open Claude and paste the prompt below with the scored theme table, current positioning, product strengths, and links or excerpts that support each theme. Ask it to draft a Google Docs memo, but require every recommendation to cite supporting call evidence. Copy the output into Google Docs and structure it around what current messaging misses, revised audience pain, 3-5 message pillars, proof points needed, objection language, competitor framing, website copy recommendations, and sales narrative recommendations. Label anything without enough evidence as a hypothesis rather than a conclusion. The memo should be written for decision-making, not as a long research report.

Share the memo with sales, customer success, support, and product using a focused review checklist. Ask reviewers whether the themes match live conversations, which phrases reps would actually use, which objections are missing, and where the memo overgeneralizes. Capture every comment in Airtable with reviewer, decision, evidence requested, and follow-up owner. Do not allow the review to become a full brand rewrite; this is a reality check on buyer evidence and commercial usefulness. The PMM should make final decisions and document which themes are approved, rejected, or marked for testing.

Convert the approved pillars into operational assets instead of leaving them in a strategy memo. Open Claude and paste the prompt below with the approved message pillars, rejected themes, buyer phrases, proof points, and usage rules from the PMM review. Ask for website hero options, product page section copy, sales talk tracks, objection responses, email snippets, ad angles, and FAQ answers. Copy the usable drafts into Airtable or Google Docs with the approved theme, source evidence, owner, status, and channel usage notes. Start rollout with the highest-impact surface, such as the main product page or the most-used sales deck, before changing every channel.

After rollout, monitor whether the approved messaging changes buyer behavior. Pull Gong tracker notes, sales feedback, reply quality, demo conversion, objection frequency, website conversion, and qualitative comments into a short measurement sheet. Open Claude and paste the prompt below with the before-and-after signals, then ask it to summarize what improved, what created confusion, and what needs another test. Copy the summary back into the positioning memo as a dated learning note and mark which messages should be kept, revised, or retired. Look for signal quality, not only volume: better questions, clearer fit, fewer repeated objections, or faster buyer understanding.