I saw a LinkedIn post recently that got me really thinking about the negative influences that could be misleading startup founders and small business owners.
This particular founder, a self-styled “solopreneur” had convinced himself his marketing was better “managed” by an AI agent (or team of agents) than by a human.
Out of curiosity, I did check out his 3 ventures — none of them showed much engagement on LinkedIn despite each sitting at a suspiciously similar number of followers, despite being vastly different companies founded years apart.
I wanted to ask him if his automation had translated to more conversions? Because I was not convinced by what I was seeing in my investigation.
If you can’t pick up what I’m putting down, I felt this guy’s whole self-image was fake (even the follower count might be explained by bots). This may come across as judgemental but I really can’t stand grifters who want to mislead others into following a path that may in fact harm their reputation.
In the hopes of avoiding the ire of the actual person who wrote the post, I got Mistral to summarise and anonymise it:
A founder built a weekend project that uses voice commands to trigger AI agents — generating images, drafting copy, and posting across platforms, all without human intervention. The stack includes interconnected AI agents, workflow automation for dozens of tasks, and real-time dashboards, all managed by a single person at zero payroll cost. The claim? With the right tools, one individual can now replicate the output of a small agency.
Founders are under a lot of pressure to deliver a huge amount of impact in a short amount of time. However, some founders are more suited for the role than others. There are some who just know which ideas are good, who to invest in, who to talk to, and how to grow. You certainly can’t replace that insight and real-world interaction with an AI Agent. And if you don’t have someone with some kind of talent on your team, you’re not going to get very far.
It also made me reflect on a project I’ve currently undertaken for another solopreneur, who is generating a decent amount of income for a single individual, but could take on staff if he could expand his business.
His work involves spending time 1-1 with his clients, outside a traditional office setting, where a lot of the work takes place via conversation, and he expresses himself best when speaking, or writing notes by hand. Despite not being glued to a computer, this is a person whose work is deeply intellectual, involving science/math, strategy, and generally complex concepts since his work makes a meaningful impact on people’s lives. He juggles a lot of different clients’ needs, and he works long hours. Spending a second shift in front of a laptop to deal with the data of it all is not appealing to him.
This is a real person I know, but his situation could apply to a lot of careers:
- Therapist
- Life coach
- Architect
- Private tutor
- Financial advisor
- Personal trainer
- Dietician
- Or a field researcher
One of the key components that remains constant across these careers is the importance of the human in the role. The types of people who are successful in these jobs build a rapport with their clients that an AI could simply not automate.
They’re also prime candidates for using AI to make their jobs better.
Here’s how I’m doing it:
A low-commitment personalised app
Using a low-code/vibe-code platform (shout out to Softr), we’re building a custom app that will create a library reference for his expertise (gathered over several interviews and from some of his existing content) as a RAG database and deliver custom feedback to clients. He still needs to approve the content (which he can do on his phone with as little as 1 tap), and can edit or change it, but it’s much faster than trying to put together spreadsheets of data himself, which is what he’s been doing.
Clients will fill out a form, their data is anonymised whenever it’s sent across an API or to an AI, and they also get access to their own tools — a “favourites” area, a “generic content” area, and periodic “update” forms to keep them on track.
If this were a profession requiring a high level of data protection, we’d also consider air-gapping the AI model to prevent any data from leaking, but as it stands, it’s a GDPR-compliant tool.
This is a fast and easy way to test out the idea as an MVP, and see if AI will really help expand his business before committing too many resources to the concept.
Augmenting his 1-1 style
Clients in Luxembourg will still be able to meet up with him in person, and clients abroad can get video packages or request a visit if he’s in their city (as he frequently travels to Italy).
This app is meant to extend his reach — and his clients’ abilities to reach him. It’s augmenting rather than replacing — at no point will the app work “by itself”, unless of course, we want to create a very low tier for less-personalised engagements. Still, that would expand his ability to hire more people (like me, as a technology partner), and probably his need, too. Other professionals in his domain could join his app, follow his unique methodology, and extend the number of users who could benefit from using it.
An ongoing project — taking feedback and expanding
This app isn’t set and forget. As AI advances, it will require refactoring and improvement. So many who think they can build a team of agents also believe they can just leave them to do their work. This is not so. Talk to any real developer and they will explain to you how spaghetti code is becoming an issue.
The other factor is that the AI model isn’t actually “there” when this professional is interacting with his clients. They may come up with new feature ideas that the AI wouldn’t be able to identify simply because it’s not listening to them the whole time. AI can really only work with the data it has. So, getting his feedback, and being one of his clients (which is how we got to know each other) means I’ll get both the customer feedback and my own user experience to help guide the future growth of this app… which, if our ambitions come true, will someday become a platform.
Can’t wait to share more, but until then, a few thoughts:
- If you’re using AI, think of where you can augment the talent that exists on your team. If you don’t have someone with that talent (such as in marketing), how are you going to verify that your agents are doing a good job?
- Think of how you prefer to work, and see if AI can enable it. You love to write long rambly notes? AI can organise it. You prefer to walk and talk (à la Steve Jobs)? AI can parse your words and organise plans and ideas for you.
- Think about how you can continuously improve your work with AI. There are advancements in AI weekly — you don’t have to stay on top of all of them, but you should stay on top of some (sign up for my newsletter if you need help). But you’re not going to be able to “create a team of AI agents” and then never think about them again. They’re going to break down. The codebase in your methodology may have dependencies that become unsupported. And all companies evolve. Even my local bakery just got a new POS system!
If you’re looking to augment your work with AI as a solo entrepreneur or a small team in Luxembourg, please get in touch.