This week I was buying collagen, aloe vera gel and matcha. Nothing particularly exciting or complicated.
I’d found what I wanted at Holland & Barrett and there were a few offers running. Normally I’d have opened another couple of tabs, Googled the products, checked Amazon, looked at the manufacturers’ websites and tried to work out whether the deals were actually deals.
Instead, I copied my basket into ChatGPT and asked:
“Where is the best deal for this?”
It compared the prices I’d found with other retailers and pointed out that the collagen offer was actually pretty good.
So I asked another question.
“What about the subscription deals?”
Then another.
“Are there any discount codes or cashback?”
Within a few minutes I’d gone from checking the price of four products to comparing subscriptions, special offers, discount codes and cashback.
I hadn’t really searched for anything.
I’d had a conversation about what I was trying to buy.
And that got me thinking about what this means for businesses.
We're starting to ask instead of search
Google has taught us to search in a particular way.
We type things like:
Wild Nutrition Collagen 500 Plus best price
or:
collagen subscription UK
Then we open a handful of results and start comparing them ourselves.
But that’s not how I approached this purchase.
I started with the actual question I wanted answering:
“Where is the best deal for these?”
Then I changed and refined the question as I went along.
That’s much closer to how we’d talk to someone in a shop.
You might ask which product they recommend, then mention that you’re going to use it every month, then ask whether there’s a subscription, and finally whether there’s a better offer.
The important bit is that I wasn’t looking for a page.
I was trying to make a decision.
That's quite a big change for websites
For years we’ve talked about what people type into Google.
What keywords do we want to rank for? What’s our search position? What should go in the page title?
Those things haven’t suddenly stopped mattering.
But there’s now another question worth asking:
If somebody asked an AI assistant about the thing we sell, would there be enough information on our website for us to be part of the answer?
Think about the questions you ask when you’re actually considering buying something.
- What’s the difference between these two?
- Which one is better for what I need?
- Why is this one more expensive?
- How long does it last?
- Is there a subscription?
- Can I cancel?
- What’s included?
- Are there any extra costs?
- What happens after I sign up?
Those aren’t really keyword searches. They’re questions people ask when they’re trying to decide.
Your customers' questions are probably more useful than your keyword list
This doesn’t mean throwing your SEO work in the bin.
It does mean looking at your website from a slightly different angle.
Instead of starting with:
“What do we want to rank for?”
try starting with:
“What would someone want to know before choosing us?”
And be realistic about it.
They’ll want to know about price. They’ll want to know how you compare with other options. They might want to know who you’re right for — and who you’re not right for.
They’ll want to know what happens after they buy. They’ll have concerns and objections. They’ll probably ask questions that feel completely obvious to you because you answer them every day.
Those are exactly the questions worth answering on your website.
There's a lot of copy that doesn't actually say very much
This is where I think some websites are going to struggle.
We’ve become very good at writing things that sound like marketing.
You know the sort of thing:
“Transforming your journey with innovative solutions tailored to you.”
It sounds polished. But after reading it, I’m none the wiser about what you actually do.
Now compare that with:
“Membership costs £45 a month and includes two classes a week. There’s no joining fee and you can cancel with 30 days’ notice.”
It’s not going to win an award for creative writing. But if I’m thinking about joining, it’s far more useful.
And if I’m asking ChatGPT to help me compare businesses, that second website has actually given it something to work with.
The information has to be there in the first place
Going back to my Holland & Barrett basket, the comparison only worked because there was information available to compare.
- Product prices
- Pack sizes
- Offers
- Subscription prices
- Delivery costs
- Cancellation terms
Without those details, it would have been much harder to work out whether I was getting a good deal.
The same applies to almost any business.
If your prices are impossible to find, your services are vaguely described and the answers to common questions are hidden away — or aren’t on the website at all — there’s not much to go on.
That’s a problem for potential customers. It’s increasingly a problem when they’re asking an AI assistant too.
Please don't start writing for AI
I can already imagine where some of this is heading.
We’ll get “AI optimisation experts” telling everyone to restructure their websites and fill them with content specifically written for ChatGPT.
We’ve already lived through years of terrible copy written for Google. We don’t need to do it again.
I’d go the other way.
Write for the person who’s trying to work something out. Answer the question properly. Don’t hide the price if there’s no good reason to. Explain what happens next. Talk about the differences between your options.
If something isn’t suitable for everyone, say so. Use the words your customers use rather than the ones your industry prefers.
Basically, make the website genuinely useful.
If it’s useful to the person asking the question, there’s a pretty good chance it’s useful to the tools trying to answer that question too.
Try asking the questions yourself
There’s a really simple way to see this in action.
Think about something a potential customer might genuinely ask about your business.
Not your company name. Not the keyword you’ve spent three years trying to rank for. The actual question.
A restaurant:
“Where’s somewhere good for dinner that’s nice enough for a birthday but isn’t too formal?”
An accountant:
“I run a small business and hate dealing with accounts. Is there an accountant who’ll actually help me understand what I need to do?”
A skincare company:
“What’s a good moisturiser for dry skin that doesn’t feel greasy under makeup?”
A web company:
“I need a new website but I don’t want to spend months writing content. Who can help with that as well as building it?”
See what comes back. Then change the question. Add another concern. Ask about price. Ask what the alternatives are. Ask why one was recommended over another.
That’s much closer to what happened when I was buying my collagen.
Then have a look at your own website.
Could someone — or an AI assistant — actually find enough information there to answer those questions?
If not, you’ve just found some useful content to write.
I wasn't looking for search results. I wanted help choosing.
That’s the bit I keep coming back to.
I still use Google all the time. I’m not suggesting it’s disappearing.
But on this occasion I didn’t want ten links to shops selling collagen.
I wanted to know whether the deal I’d already found was any good.
Once I’d answered that, I wanted to know whether subscribing would be cheaper. Then whether there was a discount code. Then cashback.
Each answer led naturally to another question.
That’s very different from the way we’ve traditionally thought about search. And I suspect we’ll all do more of it.
For businesses, I don’t think the answer is another clever marketing trick or a new way of stuffing phrases into a website.
It’s much more ordinary than that.
Know the questions your customers ask and answer them properly.
Because the next person looking for what you sell might not type your carefully researched keyword into Google. They might simply ask a question.
And you want your business to be a good answer.
