5 things to vibe code with Claude to build a better business now
getty
You have a list of things your business needs and no way to get them made. An interactive lead magnet. A better landing page to send people from Instagram. The tool you promised in your book and never built. Each one has sat there for months because building it meant finding a developer and a gap in your year.
I founded a social media agency and ran it for 10 years. Every product idea we had in those years stopped at the same point, which was the quote from the development agency. We were 20 people who could sell and write, and not one of us could build the thing we kept describing to each other.
That constraint is gone. I build with Claude Cowork now, describing what I want in English and getting something working the same day. But it’s not just me. 63% of users of these tools have never worked as a developer. The founders using them well are strengthening the three things nobody can copy, which are distribution, community and personal brand. Here are the five builds to make first.
What to vibe code before you seek external helpThe page for every platform you post on
Somebody reads your LinkedIn post, taps your profile and gets a website written for a different audience. Your homepage speaks to clients who already know what you do. The person clicking through from a post wants the thing you just described, and what they get is a menu of services and a contact form. Each platform needs a dedicated landing page.
The people coming from LinkedIn see the perfect page for them. The people coming from Instagram get a different one. Two different starting points, two different first sentences. Take your best performing post and describe the page that should sit at the end of it. Claude will have that page for you before you finish overthinking the headline.
The portal your community logs into
Run a cohort or a membership and you know the problem. Resources in a Drive folder, calls and replays scattered across email threads, and every new member asking the same questions in their first week because nothing has a home.
I built the member portal for my personal brand cohort because I wanted one place people could log into and find everything, in the order somebody would use it. That build would have been a quote and a three month wait two years ago. It took an afternoon, and the people inside it stopped emailing me to ask where things were.
The tool your customers already built badly
Customers bought your product, but they’re still struggling with an adjacent problem, getting lost in a bunch of spreadsheets and broken automations. They’ve tried vibe coding, but they don’t know how to do it, and that’s your business opportunity.
Go beyond building to solve the problems you already solve. Instead, figure out what your customers are thinking about and doing, and build for that. Your next business unlock could be in a brand new vibe coded offer.
The companion to the thing you already sell
You wrote the book and gave the talk. The person who finished it was ready to do something that evening, and what you gave them was a suggestion to reread chapter four. Every information-based product you sell has a gap between finishing it and implementing it, and that gap is where people stop. Your job is to close that gap.
My book now has a prompts playbook attached to it so readers can act on it the day they close the cover. Building that took less time than writing a single chapter, and the reader gets tonnes more value and is more familiar with my work. Whatever you already sell has a companion you have described to people a hundred times and never made, and you can describe it to a tool right now instead.
The internal job you were going to hire for
The report you rebuild every month, the onboarding you run by hand. That is the admin that made you open a job description last week and close it three lines in. None of it needs a person, and all of it needs a system you keep telling yourself you have no time to build.
Build the system first, then hire for whatever survives it. When you do hire, hire people who can use AI or want to learn it, and not be fighting to make them adapt with the times. Build one thing, use it, then start the next. Attempting all five in a week is how you burn out.
Use vibe coding to stay lean with a small team
Products get copied within weeks now, and the price of building one has fallen to almost nothing for everybody. What survives that is distribution, community and personal brand, and each of these five builds sits inside one of them.
Start with the one you have described to people the most times. Have it working by the weekend, use it badly for a fortnight, then fix the parts that annoyed you. The job description can wait until you find out how much of that role you can build yourself.
Get my free AI playbook for ambitious founders looking to scale.