Why AI Is Hard
A truthful case for slowing down and doing the hard yards with AI.
Dear reader,
The loudest voices in the AI online space are mostly talking to themselves.
If you spend any time on X or YouTube, you’ll see endless tutorials on how to improve your solo business or ways to use AI in your personal life.
It all looks incredibly easy because these demos are built for people who just don’t have standards. They don’t have brand guidelines, sign-off chains, or big clients to worry about.
But when you move away from personal use cases and into the reality of a team, implementing AI gets a lot harder than people let on.
The hidden gap between solo demos and team reality
I was working on a presentation for a client this week, trying to figure out how to convert their presentation notes into automated slides while strictly adhering to their brand template.
It is really hard to do.
Most PowerPoint templates were created pre-AI, so it's hard to stick strictly to them when recreating them in an AI tool. They mostly don’t work well. And I’ve spent hours looking into them.
The better way is to be locked into a specific ecosystem like Claude design or Gamma and redo your brand template, so it works smoothly with AI from the start.
That means compromising any precious brand templates for the sake of machine efficiency.
That is, you can create something fairly on-brand using AI, but you can’t create something STRICTLY on-brand without it being incredibly slow and cumbersome.
At that point, you might as well just write it yourself. This is where the hype hits a wall. The people winning over audiences with these “magic” tools often have FAR lower taste and don’t understand the intricacies of working with different personality types within a company or adhering to brand design systems.
The AI hype content out there is very mismatched with the reality of professional standards.
Why your hesitation to adopt AI is actually a rational instinct
This is hard for me to say as a guy selling AI services. But it’s important to be upfront.
AI isn’t ready for big teams in the way people claim. It can be incredible, but nothing will work perfectly on day 1.
Within the space I work in, it’s a lot more complicated than a one-click solution. This is why I see so many projects dragging on and on. Clients are cautious about really getting going.
They feel they should do something, and they might even put some test budget towards AI, but they don’t feel the urgency to jump in like they could because they haven’t experienced the payoff yet.
And they’re not wrong. They’re smart. They know it’s going to be harder than the big promise, and to an extent, their instincts are right.
But the problem is that this caution can lead to doing nothing at all, which has its own MUCH higher cost. No one should feel bad about being sceptical of advice written by people who have never worked inside a real business at any senior level.
A lot of these “free” automations and prompts are written by juniors or kids in their parents’ basements with almost no experience. They’re often made for people who aren’t sophisticated and have no team to answer to.
The solution: The strategy of workflow stacking
Instead of trying to transform everything at once, we have to be strategically selective. As soon as you start unpacking what can be done, more complexity arises because the work is incredibly granular. You have to pick your highest value workflows and be prepared to put in the time to get them right.
I’ve seen too many pilot projects in progress at the same time. When you have too many workflows you want to change at once, they all tend to collapse. You just need to do one well at a time.
I think of it as “workflow stacking.”
You build it like Lego bricks, one bit at a time, and stack the next success on top of the last one.
Putting in the hard yards for an AI system that lasts
I recently built a blog system for an agency that they now use on every single day. That didn’t happen overnight. It took a bit of time to work, to work for their expertise, and we had to put in the hard yards without expecting a short-term fix.
They were prepared to reinvent their ways of working for that specific result.
Being “choiceful” with what you focus on is the only way forward.
You have to decide to cut off the idea of transforming the whole company in one go. Focus on the few key things that will give you the biggest bang for your buck. It’s about deciding what actually matters to growth and ignoring the rest of the noise.
One workflow at a time.
Best,
Simon