Turn Every Call Into a Pain Point Library.
Your sales and client calls already tell you what buyers are stuck on. Meetily records them free on your Mac, and Claude files every objection into Notion, ready to script.
Phone is optional. By adding it you agree Orgrowth may call or text you about this; message and data rates may apply, and you can opt out any time.
You'll also receive my content strategy newsletter with more free breakdowns.
The Pain Point Library setup
Install Meetily on your Mac
Free and open sourceMeetily is a free, open-source meeting recorder with over 30,000 stars on GitHub. It records your microphone and your Mac's system audio together, so it works with whichever call app you use, and no bot ever joins the call. Download the free Community Edition, open the .dmg, and drag Meetily into Applications. The Mac build needs an Apple Silicon Mac (M1 or newer); there is a Windows installer on the same page. If you use Homebrew, the command below does the same thing.
brew install meetilyAllow audio and download the transcription model
Open Meetily and allow microphone access when macOS asks, then allow system audio recording so it hears the other side of the call. If you missed a prompt, open System Settings, then Privacy & Security, and switch Meetily on under Microphone and under Screen & System Audio Recording, then restart Meetily. Next it asks you to download a transcription model; keep the default Parakeet model, which is built for fast live transcription. It is a one-off download, and from then on transcription runs on your Mac.
If your own voice shows up in the transcript but the other person does not, update macOS. Meetily captures system audio with a macOS feature added in macOS 14.4.
Record your next call
Join your call as normal in Zoom, Google Meet or anything else, then press record in Meetily. The transcript fills in live as you talk, and it is saved on your Mac when you stop. Summaries are optional: under Settings, then Model Settings, pick Ollama to keep everything on your Mac, or pick Claude and paste an Anthropic API key. Choosing Claude sends the transcript text to Anthropic for the summary, so pick Ollama for any call that must stay on your machine. The library below works from the transcript either way.
Tell people you are recording, and follow the recording consent rules where you and they are; some places require everyone on the call to agree. Your library will hold other people's words, so keep your Notion workspace private.
Build the Pain Point Library in Notion with one prompt
Open claude.ai, go to Settings, then Connectors, and connect Notion, approving access to the workspace you want the library in. Then start a new chat and paste the prompt below. Claude creates the whole database for you: the properties, the six tags, and the board view that groups every row by tag.
Using my Notion connector, create a new database called Pain Point Library in my Notion workspace. Give it exactly these properties:
1. Pain Point (title)
2. Verbatim Quote (text)
3. JTBD Tag (select, with exactly these six options: Pain / Push, Trigger, Belief / Identity, Objection / Anxiety, Attempted Solution / Habit, Buying Motivation)
4. Who Said It (text)
5. Call Type (select, with the options Sales call and Client call)
6. Call Date (date)
7. Call Reference (text)
8. Used In Content (checkbox)
Then add a board view called Patterns by JTBD Tag, grouped by JTBD Tag. Do not add any sample rows. When you are done, give me the link to the database.The Patterns by JTBD Tag board is the view that pays off. After a handful of calls, the objections and triggers that keep coming up stack into the same column.
Learn the six tags
Every pain point gets exactly one tag. These six are the method, so do not rename them.
- ▸Pain / Push: the problem making their business worse right now, in their words.
- ▸Trigger: the event that made them book the call. A launch that flopped, a team member who quit, a slow quarter.
- ▸Belief / Identity: how they see themselves and their market. I am not a content person. Posting is for influencers.
- ▸Objection / Anxiety: the reason they hesitate. Price, time, been burned by an agency before.
- ▸Attempted Solution / Habit: what they rely on instead. Usually referrals, or posting whenever they find the time.
- ▸Buying Motivation: the outcome they actually want. Booked calls without sending cold DMs.
Create a Claude Project with the mining prompt
In Claude, create a Project called Pain Point Miner and paste the prompt below into the Project instructions. These rules are what keep the library clean: real quotes only, one pain per row, always one of the six tags, and never the same call filed twice.
You are mining a recorded call between me (the founder, and the seller) and a prospect or client. Extract intelligence about THEM, never about me.
For each transcript I paste, work out their name, whether it was a sales call or a client call, a three sentence call summary, and 5 to 8 pain points. Each pain point needs a short title, a JTBD tag, and a verbatim quote. The tag must be exactly one of: Pain / Push, Trigger, Belief / Identity, Objection / Anxiety, Attempted Solution / Habit, Buying Motivation.
Rules:
1. Every verbatim quote must be their exact words copied from the transcript, under 40 words. Never invent, blend or paraphrase inside a quote; leave the quote out rather than approximate.
2. The transcript may not label who is speaking. Work it out from context: I am the one explaining the offer and asking the questions. If you cannot tell who said a line, do not quote it.
3. One specific pain, belief, objection or motivation per row, with a title under 18 words.
Then file the results into my Pain Point Library in Notion: one row per pain point, filling Pain Point, Verbatim Quote, JTBD Tag, Who Said It, Call Type, Call Date and Call Reference. Before writing anything, search the library for rows with the same Call Reference; if any exist, stop and tell me that call is already filed. Finish by showing me the call summary and the rows you added.Paste each transcript after the call
When a call ends, open it in Meetily and click Copy Transcript. Open your Pain Point Miner project, start a new chat, and type one line on top: their name, sales or client call, the date, and the meeting title from Meetily as the call reference. Paste the transcript under it and send. Claude mines the call and files tagged rows straight into your library. Then spend five minutes on review: fix any tag that reads wrong and delete anything that is not a real signal. The automation drafts; you stay the editor.
Write scripts in your voice from the library
This is where the calls become content. Once you have a few calls filed, open a new chat in Claude with Notion connected, attach your voice document if you have one, and paste the prompt below. Claude finds the pains and objections that keep coming up and writes scripts built on your buyers' own words, plus an email for your newsletter.
Using my Notion connector, read my Pain Point Library. Group the rows by JTBD Tag and find the pains, objections and triggers that came up on more than one call. Ignore any row with Used In Content ticked.
Pick the three strongest patterns and write one short-form video script for each, 30 to 45 seconds spoken. For every script:
1. Open with a hook built from a real verbatim quote in the library, reworded just enough that nobody on the call could be identified.
2. Name the problem the way my buyers name it, using their words rather than marketing words.
3. Give one specific fix or reframe I would actually say on a call.
4. End with one clear call to action.
Write in my voice. If I have attached my-voice.md, follow it line by line. If not, keep sentences short and conversational, and cut any line I would not say out loud. Never use a prospect's or client's name, or any detail that could identify them.
After the scripts, write one email newsletter draft built on the single strongest pattern. Then show me which library rows each piece came from, and once I approve them, tick Used In Content on those rows.No voice document yet? My voice scriptwriter guide builds one from your own transcripts, then you attach it to this chat.
Hands-off: let Claude Code read your Meetily folder
Once the manual loop feels right, take the copying and pasting out of it. Meetily saves your meetings in a folder on your Mac (Settings shows exactly where), so Claude Code can read new transcripts straight from disk. Save the mining prompt from step 6 as a file called pain-point-miner.md, connect Notion to Claude Code with the first command below, then paste the second prompt with your folder path filled in. Stay manual for your first week though; you will tune the mining prompt to your calls.
claude mcp add -t http notion https://mcp.notion.com/mcpRead pain-point-miner.md, then read every meeting transcript in [paste your Meetily recordings folder path]. For each meeting whose title and date do not already appear as a Call Reference in my Pain Point Library in Notion, mine it using the rules in pain-point-miner.md and file the rows. Skip any meeting shorter than 10 minutes. When you are done, list the meetings you filed and the meetings you skipped.