So most people use Claude to write stuff. Sales pages, emails, an ebook nobody asked for. That’s fine, but I think the bigger win is using it to read stuff, like your call transcripts, your DM replies and your weekly numbers, and having it tell you what’s broken.
When I started in fitness I spent about a year building a $47 program called Rapid Muscle System before I knew if anyone even wanted it. It made me about $2,000 in two years. Pretty pathetic lol. If I did it again today I’d sell first, deliver it live, and let Claude chew on everything that comes back.
Quick answer: How can I use Claude as a business analyst once customers start buying my product?
Once people start buying, I would feed Claude everything: coaching call transcripts, DMs, sales calls and weekly numbers. It finds the questions clients keep asking, the openers that get replies and where sales stall. Run 10 or 20 coaching calls through it first, then all of them together, and turn what clients say into training, FAQs and ads.
- Sell a short program and deliver it live on weekly Zoom calls before you build 100 videos.
- Run 10 or 20 coaching transcripts through Claude, then all of them together for a final pass.
- Review 10 sales call transcripts to find where trust drops and which objections repeat.
- Ask for 30 ad ideas grouped by pain, results, proof, curiosity and objections, using only real customer data.
- Give Claude your weekly traffic, leads, bookings, shows, closes, cost per client, refunds, retention and profit.
Should you sell it before you build the whole thing?
You don’t need 100 videos to start. I’d ask Claude for the shortest path that takes someone from stuck to the result they’re paying for, broken into a few modules, worksheets and action steps. Then I’d go sell it and deliver the first version live on weekly Zoom calls or group coaching.
Basically you teach it while you build it. The first clients show you what the finished product needs, so you’re not sitting there guessing.
What should you do with every coaching call transcript?
After each live session I’d drop the transcript into Claude and ask it a few things. What questions came up the most, where people got confused, which explanations got someone unstuck, and what should go into the permanent curriculum.
Do that for 10 or 20 calls and then give it all of them at once. Now you’ve got the bones of recorded trainings, FAQs, templates, scripts and a troubleshooting guide, all built around problems your clients actually had. That’s way better than disappearing for 12 months and building something nobody wanted.
Let it read your DMs and sales calls
When I started I didn’t have some giant audience, I just messaged people. The example I used to teach was something like 100 DMs, 10 to 15% reply, that turns into maybe three sales calls and one client. Those were example numbers, but the point is it was measurable.
If replies are bad, fix the opener. If people reply but nobody books, fix the conversation. If they book and don’t buy, fix the sales process.
What I’d do now is paste 100 outreach conversations into Claude and say, sort these replies, show me which openers start the best conversations, list the objections that keep coming up, show me where the conversations die, and give me the next three tests to run. Same thing with sales calls. Give it 10 transcripts and ask where trust drops, what people ask right before they buy and why you lost the last few deals.
Your sales guy still has to close. Claude just helps him learn way faster than going through 10 calls by hand.
How do you turn what clients say into ads?
Once there’s some cash coming in from outreach, then I’d move to paid traffic. I’d give Claude the winning DMs, the best sales calls, testimonials and objections, and ask for 30 ad ideas split up by pain, results, proof, curiosity, objections and so on. And I’d tell it straight up, don’t make up claims or results, only use what’s in the customer data I gave you.
Take whichever ad does best and ask for 20 new hooks on the same angle without reusing the words. Then that same idea can become a YouTube ad, a landing page headline, an email and a follow up sequence. One offer, a bunch of ways to say it.
Which numbers should you give it every week?
This is the part I think most people skip. Every week I’d hand Claude my numbers, so traffic, leads, how many book, how many show up, how many close, what a client costs me, refunds, retention and profit, and ask it what the biggest bottleneck is.
If traffic doubled and revenue barely moved, it’s a conversion problem. If sales went up and profit went down, go look at ad cost, delivery cost or refunds. This is why we look at the numbers. The numbers tell us what to do.
Now, Claude can’t pick your market for you, it can’t fake experience you don’t have, and it can’t save an offer nobody wants. You still have to know your stuff and go sell it. It just makes the boring parts way faster.
I go through the whole model, the pricing and the funnel math in the full breakdown on tannerchidester.com.
Frequently asked questions
How do I use Claude to analyze coaching call transcripts?
Run each live session’s transcript through Claude and ask for the most common questions, where people got confused and which explanations landed. Do 10 or 20 calls one at a time, then give it all of them together for a final pass. That turns into training material, FAQs, templates and a troubleshooting guide built from real client problems.
How can Claude tell me why my DM outreach is not converting?
Treat outreach like a funnel and give Claude your DMs and sales calls. Say 100 DMs get a 10 to 15% reply rate, about three sales calls and one client. Poor replies point to the opener. Replies with no bookings point to the conversation. Bookings with no purchases point to the sales process.
How do I write ads with Claude from customer language?
Once outreach is making money, give Claude your winning messages, testimonials and objections. Ask for 30 ad ideas grouped by pain, results, proof, curiosity and objections, then 20 new hooks from the best angle. Tell it to use only claims and results from the data you gave it. No invented claims.
What weekly numbers should I give Claude to find my business bottleneck?
Traffic, leads, bookings, show ups, closes, cost per client, refunds, retention and profit. Then ask for the main bottleneck and three next tests. If traffic doubles and revenue stays flat, it is conversion. If sales go up and profit goes down, look at ad cost, delivery cost or refunds.
Should I build my whole course before I sell it?
No. I spent about a year building Rapid Muscle System, a $47 program, before checking if anyone wanted it. It made about $2,000 over two years. Sell a short, result focused program first, deliver it live on weekly Zoom calls or group coaching, and build the full product from what those first clients need.
Want to know what’s capping your business right now? Take my free 90 Day Ceiling Plan. Answer 18 questions and get a written plan, built on your own numbers, that names the one thing holding you back and what to fix first.
Written by Tanner Chidester, founder and CEO of Elite CEOs. He has done over $150M in sales, worked with 50,000+ clients, and hosts the Infinite Income Podcast.









