A few weeks ago I built a carousel generator over the weekend.
It takes a podcast transcript and turns it into a full set of social media images — the kind that would normally take a designer a few hours to put together. This one runs in minutes.
I built it for Asian Efficiency. We record a podcast almost every week, and I had been wanting a way to turn those episodes into visual content without adding hours of work to someone's plate. So I sat down, built an AI agent in Lindy, tested it on a recent episode, and it worked.
Here is what I did not fully anticipate: that thing is now one of my best consulting demos.
The Problem With Traditional Case Studies
Most consultants and service providers build case studies the slow way. A client hires you, you do the work, you document the results, you clean it up, you put it in a deck. Then you bring that deck to the next meeting.
There is nothing wrong with that. But there is a delay built in. You have to do client work before you can demonstrate client results. And by the time the case study is ready, the example might already feel dated.
There is a faster loop.
Build It for Yourself First
When you solve a problem in your own business, you get two things at once: the tool and the proof.
I know the carousel generator works because I run it on my own content. I can show someone the actual output from last week's episode. When a potential client who creates a lot of content asks “can this work for my situation,” the answer is not theoretical — I can pull up something real and run it live.
That demo lands differently than a case study. A case study says “I helped someone else do this.” A live tool says “I use this myself and here is what it does right now.”
This is something I have noticed across a lot of the consultants doing well in AI implementation: their own operations are their R&D lab. They do not separate “building the business” from “solving the problem.” They solve the problem in their own workflow first, and the client work follows naturally from what they have already proven works.
The Feedback Loop Is Real
It runs in both directions, which is the part worth paying attention to.
The tools I build for Asian Efficiency become demos for consulting clients. And the problems clients bring to me teach me what to build next for AE.
Working with a real estate firm on meeting automation gave me ideas for how to handle my own post-meeting workflows. A session with a CPA helped me think about email filtering in a way I had not before. I came back from those engagements and built things for my own system that I now use every day.
Your client work is also R&D. You are just not logging it that way.
One consultant I talked to recently described building a talent pool from his workshop attendees — people who came to learn would then take on overflow client work. His teaching lab became his workforce. Same principle: what you build for one purpose feeds the other.
What This Means Practically
If you do any kind of consulting, coaching, or service work, there is a simple question worth asking at the start of any new project: would this same solution be useful in my own business?
Not always. Sometimes a client's situation is specific enough that the solution does not generalize. But more often than not, if the problem is worth solving for them, it is worth solving for yourself too.
Build it for yourself first. Document it as you go. When it works, you have already done the proof-of-concept work. Then you bring it to clients as something you use, not just something you have theorized about.
Your business stops being overhead. It becomes the product.
If you want to see what this looks like in practice: I do a small number of one-day AI workshops in Austin and occasionally other cities — working through real workflows live, so you leave with something you have built yourself. If that interests you, the best starting point is the Productivity Academy — I post workshop dates there first.
