This post was inspired by a recent webinar we hosted with AI strategist Tristan Day and Workiro President Dave Owen. It wasn’t a product pitch. It was a frank, 60-minute conversation about why so many AI projects fail before they start, and what real AI readiness looks like in the messy reality of day-to-day business.
Let’s start with something uncomfortable: most businesses aren’t remotely ready for AI.
You might have a ChatGPT tab open right now. Maybe someone in marketing is proudly pasting prompts into it. And look, we get it. It feels like progress. But it isn’t.
Here’s what Tristan Day, founder of DemosAI, said in our latest webinar: “This is the first time where we’re seeing digital transformation being driven by FOMO rather than planning.”
That’s your red flag.
So, if you don’t want your AI project to turn into yet another expensive shelfware experiment, here’s how to lay the groundwork.
1. Stop thinking about the tech
We know. It’s shiny. It talks. It writes emails. But AI is a tool, not a magic trick. And tools only work when they’re aimed at a real problem.
Start here: what are the actual business problems you want to solve? (Not the ones you think sound strategic. The real ones. The messy, unglamorous ones that cost you time and money every day.)
As Tristan puts it: “The advantage isn’t going to be having the tool. It’s how you use the tool.”
2. Audit your data like your job depends on it
Because it kind of does.
85% of AI projects fail due to bad data, according to Gartner. Not bad models. Not a lack of budget. Bad data.
And the worst part? You probably don’t even realise your data’s that bad until the AI starts spitting out answers that make zero sense.
Dave Owen, President of Workiro, summed it up best: “AI can only work with what you provide. If your data’s fragmented, siloed, or buried in inboxes, your results are going to be equally fragmented, siloed, and buried in nonsense.”
So go look in your ERP. Your CRM. Your document management system. Ask yourself:
- Can I find what I need in under two minutes?
- Is this the same version sales is using?
- Is anything labelled ‘final_v3_FINAL-revised.pdf’?
If yes, you have work to do.
3. Build around people, not platforms

Technology doesn’t fail. Adoption does.
You could buy the world’s smartest AI tomorrow and it would still fall flat if your teams don’t trust it, use it, or even know what it’s for.
“If you're not solving the problem people actually have, you’re going to get eye rolls instead of adoption. And rebuilding trust takes ten times longer than building it.” says Tristan.
Bring people in early. Give them ownership. Make the benefit obvious.
If it doesn’t make their working life better, they’ll ignore it. Or worse: actively resist it.
4. Fix the plumbing before installing the smart mirror
Yes, it’s a metaphor. But it’s also your reality.
If your ERP, CRM, and document system don’t talk to each other, AI can’t do anything useful. It can’t read minds. It reads APIs.
As Dave puts it, “We’re talking about systems here like ERP, CRM, DMS. They are the beating hearts of your business operations. If they’re not integrated, your AI doesn’t stand a chance.”
Want to know how AI-ready your business actually is?

We built a free scorecard to help you figure it out in 60 seconds flat. No waffle. No jargon. Just a clear snapshot of where you stand across data, systems, culture, and more.
👉 Take the AI Readiness Scorecard now
Sign up for the full session with Tristan and Dave. It might just save you from another doomed AI pilot.