The 'AI' label tells you nothing. Make them prove it.
Every product now says 'AI', and the label tells you nothing about whether it can actually change how your business runs. There's a five-minute demo test that sorts the real thing from a chatbot bolted onto last year's software.
We spent two hours last week setting up a tool that calls itself "AI-first." When we finally asked its AI to do something useful, like set up a simple workflow for us, it handed us a link to the help centre. It could answer questions about the product; it couldn't change anything in it. The label was doing a lot of work the software couldn't back up.
"AI" is now stamped on almost every product you'll be sold. The word has stopped meaning anything, because everyone uses it and nobody has to prove it. So the label tells you nothing. The only thing worth knowing is whether the tool can actually change how your business runs, and the only way to find out is to make them show you.
AI-first versus AI-native
The gap is easy to describe. An "AI-first" tool, most of the time, is a chatbot bolted onto software that was built the old way; it can tell you how to do something, then leaves you to click through the menus and set it up yourself. An "AI-native" tool is one you can simply instruct: you tell it what you want, say "when a new enquiry comes in, create the contact and start the follow-up sequence", and it does the work.
You don't need to care which architecture is underneath. You need to care about the difference you'll feel: can the people who run your business make the tool do things by asking it, or do you still need a specialist to translate every request into the software's internal logic first? One saves you a few minutes. The other changes who in your business can actually get things done.
The test is a demo, not the sales pitch
The way to cut through it takes about five minutes, and it costs you nothing. Don't argue about the technology; ask them to do a real task, live, by instruction. "Set up this workflow. Move this record. Build this rule, by telling your AI to do it, while we watch."
A genuinely AI-native tool does it in front of you. A paint job stalls: it points you back to the manual, or the demo quietly cuts to a person on their side doing it by hand. You will know which one you're looking at almost immediately, and no amount of polished marketing survives the request. The demo is the truth serum the label isn't.
What you're actually buying
When a tool can be run by instruction, your own team configures and operates it in plain language — no specialist sitting in the middle, no month of learning where every setting hides. That's the thing worth paying an "AI" premium for. A chatbot sitting on top of the same old menus is not; it's last year's software with a new sticker, sold at this year's price.
Both will say "AI-first" on the website. Both will demo beautifully if you let them run the script. So don't. Ask the one question the label can't answer for you, can it actually change how my business runs?, and make them prove it, live, before you sign anything.
Weighing up an "AI" tool and not sure there's substance behind the label? Book a call and we'll put it through the demo test with you before you commit.
Robin Carswell
More on
How to audit your AI dependency
Would your business still function if OpenAI went dark tomorrow? Most founders can't answer that with confidence. Here's the audit that surfaces it.
AI tool or AI workflow: the difference that matters
When two founders say they 'use AI in the business', one may be pasting into ChatGPT and the other may have it wired into their pipeline. Same technology, very different dependency.
The AI features you didn't sign up for
Every SaaS tool you pay for is quietly shipping AI features by default. You didn't procure them, you can't cleanly turn them off, and your team is now building workflows around them.