A real-world breakdown of why AI-generated system instructions often fail in practice, and why understanding your business tech stack is essential for successful integrations and automation.

AI vs Business Systems: Why AI-Generated Tech Advice Often Breaks in Real-World Integrations

I spent almost 3 hours trying to get a code to work this afternoon so we can get a Kajabi checkout to send email data to Jotform without the customer having to enter their email again.

I was sent instructions with tags and codes I could supposedly set up. AI had clearly made these suggestions... except it didn’t have the full picture.

We’re not sending people directly to a Jotform link because we wanted the form to be embedded in a Kajabi thank you page to ensure it’s aligned with our branding.

So when I went to follow the instructions, they were already incorrect. It was telling me to set up a redirect URL but then that would need to override the custom thank you page we’re already using, which is not a viable solution.

So then, I spoke with the AI bots on both Jotform and Kajabi to see if there was another workaround.

When AI Instructions Don’t Match Your Actual System Setup

I was advised to create a hidden field and then copy a code into the Kajabi page where I already have the form embedded and that should map the email data across after checkout.

Sounded promising, so I copied the codes... made various edits. And kept testing.

First the data wasn’t transferring across. Then it showed on the form but instead of actually inputting the email, it just kept adding {customer email} on the form.. every time I asked, I was given a different variation of the code to enter.

Sometimes it completely deleted the form embed, sometimes the email field would show up on the form submission, other times it wouldn’t... not once did my email actually map across though.

After all the different tests and revisions, I finally got hold of a human support person who confirmed what I was trying to do isn’t actually possible between these platforms.

AI bots on both platforms said it was. AI gave my client instructions which sounded legitimate, she passed them on to me, I attempted to follow them.

But AI only had half the picture. And herein lies the problem...

• AI tools are not all-knowing.

• They are not inherently accurate.

• And sometimes they are confidently generating instructions for scenarios that are not the scenario you actually have.

The Problem With Blindly Trusting AI for Business Systems

This is where people can get themselves into trouble.

Because if you don’t deeply understand systems and tech, you’ll think: “Well AI told me this was possible…”

So you paste random code into platforms, implement workflows you don’t fully understand, skip proper testing, and accidentally create broken systems because the instruction sounded convincing enough to trust.

But just like most things... systems don't work in isolation, they work relationally. That's why getting them to talk together is called INTEGRATION..

One platform’s limitations affect another platform. One workflow decision impacts the entire customer journey.

One “small tweak” can completely break the logic of how data is communicating behind the scenes.

Why This Proves AI Won’t Replace Systems Experts

At least this proves I don’t really need to be worried about AI replacing my work any time soon.

And all the people banging on about how you’ll be left behind if you don’t adopt it, clearly don’t really understand how humans work.

Having information doesn’t mean you know what to do with it.

In the time it takes someone who isn’t tech savvy to try and figure out what AI actually wants them to do... or why it’s not working... they’re gonna end up getting annoyed and wanting to outsource it to a human.

That’s where my work becomes valuable.. become it comes from understanding:

• how systems actually communicate

• what’s technically possible

• where bottlenecks happen

• which platforms fit the bigger vision

• and how to build things properly from the beginning

Kajabi vs Ivorey: Choosing the Right Platform for Complex Business Systems

What’s happened in this situation is a really good example of why strategy and implementation need to work together.

The strategist I’m working with is brilliant at vision and ideas. The strategy is solid.

But the execution is hitting hurdles because what’s being asked for isn’t actually possible within the chosen platform.

We’re constantly having to come up with workarounds because of Kajabi’s limitations, which in the long term are not going to be ideal as business volume expands.

And had I known the full long-term vision from the start, I would’ve been able to say I don’t think Kajabi is the best platform for this build.

I think Ivorey would make far more sense in this model because the forms, automations, workflows, and communication between systems would all be able to exist within the same platform. Making the communication far more seamless, and the steps for all involved would be dramatically reduced.

Now to be clear, there are situations where I would still absolutely recommend Kajabi over Ivorey.

Just because I decided to move away from it, doesn’t mean I don’t still see what a great platform it is.

Especially for course, coaching, and membership models.

But with a business that has a lot more moving pieces and needs much more complex workflows, I am quickly learning how limited Kajabi actually is.

This isn’t about one platform being universally “better.” It’s about choosing the right infrastructure for the vision you’re trying to build.

But, since I'm an Ivorey affiliate I wouldn't miss an opportunity to give it a little plug. 😜 If you would like to try Ivorey for yourself, you can here.

Why Systems Diagnostic Sessions Are the First Step in My Work

And this is exactly why my Systems Diagnostic Sessions are now the starting point for every client.

Because most people come in thinking they need:

• a funnel

• a workflow

• an automation

• a new platform

• a piece of tech

But often the real issue is:

• unclear ecosystem design

• mismatched platforms

• bottlenecks they can’t yet see

• or trying to build advanced functionality on systems that weren’t designed for it

My job is to zoom out and see the bigger vision and what platforms are going to be able to hold that now and into the future.

To look at:

• Where is the business going?

• What needs to communicate together?

• What are you trying to build long term?

• What systems actually support that vision?

And from there we decide:

• Do you implement it yourself?

• Do I consult?

• Do I build it for you?

AI Is Only as Good as the System Thinking Behind It

Because most people discover very quickly that systems are a lot more nuanced than “just ask AI.”

The future doesn’t belong to the people who blindly use AI.

It belongs to the people who know how to think critically, build intentionally, and understand how systems actually work.

Yes, there is absolutely a place for AI within that. But it only ever reflects the understanding of the person using it.

No understanding = no leverage.

Buy into the hype all you want, but at the end of the day AI output is only ever gonna be as good as the input. So without a basic understanding of how specific systems work and communicate with each other, then AI could be giving completely wrong advice.

Like anything… systems are nuanced… each scenario is different and the more complex the workflow, the more important it is to think about whether you’re building in the right platform/s to save a looooot of headaches down the line.