In partnership with
You know the problem. Now build the fix.
You've spent years close to a broken industry or process. You know exactly what would fix it. But you’ve never had the time or resources to make it happen.
With Lovable, your expertise is all you need. Describe the solution to your problem — a website, a dashboard, an internal tool — and get professional-grade software back. You make every call on how it works, because you're the one who knows what it needs to do.
Start today and take $5 off your first month.
Discount of $5 applies to first month of any paid plan. You need to sign up via this webpage in order to claim the discount. The discount applies to new and returning users without an active paid plan. Terms apply.
My favorite tools
Wispr Flow: Why type when you can talk?
Granola: Meeting transcripts connected to Claude through MCP.
Claude: My daily driver for AI. The first 3 people to sign up through this link get a free week of Pro.
Muse: Meta's new AI agent. Add code CX40H3 in your settings within 48 hours of joining to unlock 1 billion free tokens.
11:31 AM - Meeting ends
11:33 AM - I grab a snack from the kitchen
11:35 AM - Claude created 12 new ads based on ideas shared from the meeting that are ready to be launched
Let me show you how!
How many inputs do you feed your AI when you prompt it?
What we built
Granola captured meeting notes from a call discussing new ad ideas.
It connects to Claude through MCP, making transcripts super easy to leverage in prompts. (MCP = the connector that lets Claude read another tool directly.)
I had Claude use both an ad copywriting and design skill built on our last 180 days of Meta ad performance.
Along with an internal Confluence repository which houses emerging use cases from customers of the feature these ads are about.
Last, I pointed Claude to an existing Figma board where it could drop in these new ads.
Let me break it down!
Granola: the meeting notes, where the new ad ideas came up
Ad copywriting skill: hooks and lines that converted over 180 days of Meta ads
Ad design skill: the visual styles from our top performers. Want to learn more about both skills? Check out my last newsletter, where I break down how we built them.
Confluence: customer use cases, so every claim is accurate to what the feature can do
Claude: one prompt, all four sources
Figma: where the finished ads land, dropped straight into frames
And walah!
A dozen ready to publish ads with snappy copy based on best performing styles, accurate to what the feature can do, and ready for the big stage.
The cool part: those two skills are the same ones from #034, so they're re-scored every Monday as new ad performance comes in.
It's a small workflow but proves how much more powerful AI can be when you feed it more inputs and better context.
Why this matters
Better inputs = better outputs. The same prompt with four sources behind it beats a blank chat every time.
Ideas don't die in the meeting notes. What gets said on the call is ready to launch four minutes later.
Fast doesn't mean guessing. Every ad starts from the hooks and visual styles that already performed.
The claims stay accurate. Confluence keeps the copy true to what the feature actually does, so speed doesn't create cleanup for the product team.
After a meeting, what's the one thing you wish was already done by the time you got back to your desk?
Same system, different problem
Take the ads out of it and this is a pattern any revenue team can use:
Capture the conversation where the ideas come up
Add the skills that hold what's already proven to work
Add the source of truth that keeps it accurate
Point Claude at where the finished work should land
A few ways to point it at revenue:
Sales follow-ups: call notes, a follow-up skill built on emails that booked next meetings, case studies → "Draft the follow-up for today's call with the 2 case studies closest to their use case."
Proposals: discovery call notes, a proposal skill built on closed-won deals, the pricing sheet → "Draft a proposal for this prospect using the structure that's closed fastest this year."
Expansion plays: QBR notes, product usage, release notes → "Which features did they ask for that we've since shipped? Draft the upsell email."
Objection handling: this week's sales calls, the messaging skill, product docs → "Turn the objections from this week's calls into a one-pager for the team."
Cheers,
Matthew Shelton
P.S. Missed #034? Here's how we built the ad skills Claude used here.
