If you’ve read our blog before, you’ll know that writing an article about how to rank in ChatGPT isn’t really our thing. We’ve never been fans of chasing the latest algorithm update, whether it’s in Google SERPs, social media or now AI.
Our view has always been pretty simple: create genuinely useful content, look after your reputation, and you’ll generally be in a good place. Good content performs well in search engines, and the very same content is often what AI tools draw upon when answering questions.
So no, this isn’t a guide to optimising for ChatGPT.
However…
When someone asks ChatGPT a very specific question like, “Should I buy from company X?” or “What’s this product like?”, it doesn’t simply summarise the company’s website. It pulls together information from a range of different sources to build a picture of that business.
That makes it worth understanding where it’s looking for information.
Not because you should suddenly start trying to game those platforms, but because if prospective customers are asking AI about your business, it’s useful to know what information it is likely to find, whether you’re represented on those platforms at all, and what sort of impression they create.
It’s not really that different to a diligent human doing their own research – where would they look and crucially, why?
Back to my heading, I asked ChatGPT where it would typically look if someone in the UK wanted an overview of a company before buying a consumer product or service.
After some digging, it gave me a list of a few sources and then quite a helpful summary about trying to balance four different types of information.
- Official information, what the company says about itself.
- Independent verification, things like regulators and professional review sites.
- Customer experience sites and forums.
- Reputable news sources – reporting about recalls, legal cases, investigations or significant company news.
It also gave an indication of the sites it most commonly refers to. Company websites were almost always part of the picture, alongside platforms such as Reddit, Trustpilot, Google Reviews, and major retailer reviews (Amazon, John Lewis, Argos, Boots, Currys), as well as specialist professional review sites like Which? AutoCar, independent comparison and review sites like MoneySavingExpert, Byrdie and, where relevant, news sources such as FT, Times, Telegraph and sector press (both consumer and trade).
Depending on the type of question, ChatGPT may also look at trust signals such as Companies House or industry regulators including FCA, ASA, Ofcom, Ofgem, DVSA, MoneyHelper. These don’t immediately form a judgement on whether a product is good, but they can help establish whether a company is legitimate, established and properly regulated.
The most interesting part wasn’t the list itself though, it was how ChatGPT said it uses those sources. It doesn’t rely on any one website. Instead, it looks for recurring themes across multiple sources because every platform has its own strengths and weaknesses.
For example, it acknowledged that customer review sites can paint a distorted picture. People are often much more motivated to leave a review after an exceptionally good or exceptionally bad experience than they are after an entirely satisfactory one. One angry Trustpilot review doesn’t define a business, and neither do ten suspiciously enthusiastic five-star ratings. The aim is to build an overall picture rather than relying on a single opinion.
It depends what you’re buying
Perhaps the most interesting finding is how much the sources change depending on what someone is looking to buy.
For beauty products, community experience carries a lot of weight. ChatGPT said it would commonly look at places like Reddit, Boots reviews, Superdrug reviews and specialist beauty publications alongside the manufacturer’s own information. That’s because skincare and cosmetics are incredibly personal. One person’s holy grail moisturiser is another person’s allergic reaction.
For home appliances, the balance shifts. Independent testing becomes much more important, so Which?, retailer reviews, repair forums and manufacturer specifications become key sources. People want to know whether their washing machine is likely to survive beyond its warranty.
Cars are another category entirely. Here, ChatGPT leans towards sources like Auto Express, What Car?, Honest John, Parkers and manufacturer information, backed up by owner forums and Reddit.
Mortgages are different again. The Financial Conduct Authority, MoneyHelper, lender websites, MoneySavingExpert and comparison sites become important here. When you’re borrowing hundreds of thousands of pounds, regulation and product terms carry more weight than anecdotal opinions.
What does this mean for brands?
The obvious takeaway is that your own website still matters. A lot.
If ChatGPT is almost always checking a company’s own website, then making sure your product information, FAQs, policies and company information are accurate and up to date remains incredibly important.
But that’s only one piece of the puzzle.
If customers are discussing your products on Reddit, leaving reviews on Trustpilot or talking about you on retailer websites, those conversations may also influence what someone hears when they ask AI about your business.
That doesn’t mean you need to obsess over every review or suddenly sign up to every platform going.
Instead, it’s worth taking stock of the places that matter in your sector. You don’t need to be everywhere. The internet is simply too big for that, and every industry has its own trusted sources.
But you should know what these are and where your customers are likely to do their homework, because increasingly that’s where AI is doing its homework too.
So, should you start optimising for ChatGPT?
Absolutely not.
Nothing I’ve found suggests businesses should be doing anything radically different from what we’ve been recommending for years. And it certainly isn’t about publishing ‘AI-friendly’ content.
Instead, it described pulling together information from lots of different places, your own website, independent review sites, customer discussions, regulators and reputable news sources, then looking for a consistent overall picture. And that’s what AI tools should be good at – analysing a mass of data and making some sense of it.
So rather than worrying about ranking in ChatGPT, spend your time building a business that earns positive reviews, publishes genuinely useful content and maintains an accurate presence across the places that matter in your industry.
As it turns out, that’s pretty much the same advice we’d have given before ChatGPT came along.