When did you last do something because AI told you to, only to realise hours later it was wrong?
If that question makes you uncomfortable, good. It should.
I run an AI consultancy. I help small businesses use these tools properly. I have been working with AI every day for a long time. And in the last few months alone, I have wasted hours following AI instructions that were confident, detailed, and completely wrong.
Not wrong in an obvious way. Wrong in the way that only becomes clear after you’ve already done the work.
This article is about that. About what actually happens when you follow AI blindly. About the LinkedIn posts telling you to hire Claude as your SEO Executive. And about the one thing that separates people who get real value from AI from people who just get burned by it.
Two AIs. The Same Wrong Answer. Confidently.
I was building my new website. I hit a URL redirect issue. A technical problem, but not an unusual one. The kind of thing that has a simple fix if you know where to look.
I asked Claude. I got a detailed response. The solution: duplicate the entire website and point one version to the other.
That seemed extreme. So I asked Gemini.
Same answer. Duplicate the website.
Two different AI tools. Two different companies. Both telling me the same thing. If I’d been less experienced, I would have assumed that two independent sources agreeing meant it was correct. That is not how this works.
I pushed back. I told them both that duplicating a website seemed like a significant overreaction to a simple redirect problem and asked them to think again. Eventually, both arrived at the real answer. A few lines of configuration. Five minutes of work.
The original answer from both was not a lie. It would technically have worked. But it was a sledgehammer solution to a problem that needed a screwdriver. Neither AI stopped to ask whether there was a simpler way. Neither flagged that their suggestion was disproportionate. They just gave me steps that would eventually solve the problem, without asking whether it was the right approach.
I had to convince two AI tools out of the wrong answer. Think about that. And then ask yourself: would you have known to push back?
Two Hours. One Question I Should Have Asked Earlier.
A few weeks later, building a blog post. The table of contents links were not working. I asked for help.
The AI told me to go into each heading, find the HTML Anchor field, and manually type in an anchor ID. Fifteen headings. One by one.
I did it. Then it found a new problem. Go into the raw HTML, it said, and remove duplicate IDs from every list item. I was about to start.
Something stopped me. I asked: is there a simpler solution I’m missing?
Yes. Delete everything. Use the Rank Math Table of Contents block. Handles everything automatically. Takes ten seconds.
Rank Math had been installed on my site the whole time. The AI didn’t know to ask what tools I already had. I didn’t know enough to mention it. Two hours of unnecessary work because neither of us stopped to ask the right question first.
I am telling you both of these stories because I want you to understand something. These are not beginner mistakes. This is how AI actually works, even when you know what you’re doing.
Why AI Sounds Right Even When It Isn’t
There is a word that gets used in AI circles: hallucination. It sounds dramatic. In practice it is mundane and dangerous.
Hallucination is when AI states something false as if it were true. No flag. No hesitation. No “I’m not certain about this.” Just the wrong answer in the same tone as the right one.
With a human expert, you pick up signals. They pause. They say “I think” or “let me check.” Their tone shifts when they’re less confident. AI has none of those tells. It answers a question it doesn’t know the answer to in exactly the same way it answers one it does.
And here is the part the LinkedIn posts leave out: two AIs agreeing does not make something true. Both Claude and Gemini gave me the same wrong answer. AI systems can share the same training biases, the same blind spots, the same tendency to reach for complex solutions when simple ones exist. Agreement is not validation.
If you are using AI to make decisions in your business — pricing, strategy, technical, legal, financial — this is the most important thing to understand before you start.
The SEO LinkedIn Posts That Worry Me
You’ve probably seen them. “I hired Claude as my SEO Executive.” “I replaced my copywriter with AI.” “Here’s how I built an AI team of six for £0 a month.”
They do well. Thousands of likes. People share them. They feel like insider knowledge.
I understand why. The people writing them have often had a genuinely useful experience. They’re excited. That’s not dishonest.
But here is what they leave out.
The person writing them usually has enough background to catch the mistakes. They know when the AI’s SEO advice is wrong because they’ve done SEO. They know when the copy needs fixing because they’ve written copy. The AI is a fast first draft. Not an independent decision maker.
A small business owner with no SEO background reads that post and thinks: great, I’ll just hand this to AI. They are in a completely different position. They don’t have the background to catch the errors. So the errors go uncaught.
That is not a small problem. That is your website optimised for the wrong search terms for six months. That is customer emails going out with mistakes in. That is a decision made on numbers an AI got wrong.
The posts show the highlight reel. They do not show the workflow. They do not show the checking. They do not show the ten years of experience quietly running in the background, catching the things the AI gets wrong.
And Then There’s Your Precious Data
The posts that worry me most are the ones that involve business data.
“I uploaded our client list and asked AI to segment it.” “I fed our financials into ChatGPT.” “I pasted the email thread with my client and asked AI to draft a response.”
The answer here is not “never put data into AI.” That would be too blunt and would stop you from doing genuinely useful things.
The real question is: which tool, with what data, with what safeguards?
The free version of ChatGPT has different privacy terms than the paid business tier. Pasting client data into it is a GDPR question, not just a technology question. Healthcare tools built with proper encryption for clinical notes are a completely different matter to a general purpose chatbot.
The LinkedIn posts don’t make this distinction. They just say: here’s what I did. And someone reading has no way of knowing whether what they did was fine or a serious mistake.
What to Do Instead
Use AI. It is genuinely useful and I use it every day. But treat it as a capable assistant with a confidence problem, not as an expert whose output you can accept without checking.
Always ask if there’s a simpler way. Before following a set of AI instructions, ask: is there a simpler solution? What tools do I already have that could do this? That one question would have saved me both stories above.
Push back when something feels wrong. I had to convince two AI tools out of duplicating my website. If your instinct says the answer is disproportionate, question it. AI responds well to being challenged. It does not volunteer a better answer unless you ask.
Agreement is not validation. Two AIs giving you the same answer does not mean it is correct. They can share the same blind spots. Always apply your own judgement, especially on anything consequential.
Know what you’re putting into it. Before pasting anything sensitive into an AI tool, check its data policy. The tool matters. Free and paid tiers are not the same. General purpose and specialist tools are not the same.
Read the LinkedIn posts with scepticism. Not cynicism. But remember: they are showing you what worked for someone who already knew enough to catch the mistakes. That does not automatically transfer to your situation.
Slow Down. Then Use It.
The businesses getting the most from AI are not the ones moving fastest. They are the ones who slowed down long enough to understand what they were using before they used it.
AI will save you time. It will help you think. It will do things in minutes that used to take hours. All of that is true.
But it needs a human behind it who understands the problem, questions the output, and knows when to say: hang on. Is there a better way?
I’ve had to learn that the hard way. Twice. In the same month.
I’m telling you so you don’t have to.
Want to Use AI Without Getting Burned?
I run free AI workshops for North London small businesses and independent professionals. We cover exactly this — not just what AI can do, but how to use it without losing time, money, or data in the process.
If you’d rather talk one to one, a £75 introductory consultation is the place to start. No jargon. No hard sell. Just a practical conversation about your specific situation.
Get in touch at joe@joeai.co.uk.