Jenna says: “My number one tip is for AEO, and it's all about your website.

What we're seeing in AEO is somewhat different than SEO, in that you can no longer go broad with keywords. AEO is all about getting specific, and it starts with your website.”

We're talking about Answer Engine Optimization here, obviously, aren't we?

“That's correct. AEO: Answer Engine Optimization.”

For getting keywords to be very, very specific, you mentioned that it's important to build a public documentation on your site that answers very specific questions.

What sort of public documentation are you referring to?

“The way that it works is, it's first going to your website to understand who you are, what you do, who you sell to.

When I say public documentation, your website needs to answer all of those questions. Think of it as: your website is the tool that's answering what your customer might type in to an AI chatbot. Note that, when your customer goes to an AI chatbot, they're likely doing deeper research right in the chatbot than when they're going to Google and they're typing in a keyword. Those questions get much more specific.

The first step in answer engine optimization is to answer all of those questions through your website to make sure that your content is being summarised as the response to those questions.”

It's very important that it's your website, as opposed to some wiki site or some social sites out there?

“We actually have a free tool that does AI visibility, so we see a lot of websites come through and do an audit on how they're showing up in AI. We rarely see these third-party websites that are talked about in a lot of the articles around AI search – things like Reddit don't show up that often, we don't see LinkedIn show up very often, we don't see YouTube show up very often – as the top sources.

Whenever we look at an industry, we'll often see that first-party websites are the most cited sources to start. AI search is quite good at understanding what industry you're in, who are your competitors, and then how to compare you to those competitors – which is why the top sources for AI mentions within an industry are often the first-party websites.”

I'd love to get into how you're going about structuring your website, maybe with markup code or specific ways of answering questions, but in terms of what questions you need to optimize for, how do you know which questions are the key ones initially to try to optimize for?

“This is such a great question. It's hard because, unlike in the world of SEO where you have keyword data, there currently is no data – meaning none of the AI models are releasing what questions are being asked in AI. It leaves you with that task of understanding your customers, what their buying process looks like (what questions they have along that customer buying journey), and then making sure that you are answering those questions through the content on your website.

The first step is often going to Claude or ChatGPT and asking it, ‘What questions would this ICP, this customer, ask when looking for this solution?’”

Okay, brilliant: Ask AI search engines what you should be incorporating on your website in order to actually make it more likely that your website is surfaced to your ideal ICP.

Is there any search volume at all? You say that keywords aren't important here. In terms of prioritising, obviously, AI can come up with lots of different suggestions – maybe 10, 20, or 100 different suggestions of pieces of content that you could be trying to actually publish on your website or different terms that you should be optimizing for.

If volume is very difficult to find, is it a case of just going through the different options and having a think about the profile or the kind of audience that you're trying to target and selecting content based upon who you think is the ideal customer to begin with?

“Yes, it's exactly that. We're kind of back at Marketing 101. It's understanding your customer and their buying process.

The tip I would give is that the questions should be commercial, so they should be questions where an AI would respond with a recommendation for a solution. You can get quite broad if you're going informational, but note that an AI responding to an informational question is not going to give a recommendation, which means it's unlikely that your brand will be mentioned in the answers.

You want to make sure that the content you're developing is answering the commercial questions: ‘What is the best tool for XYZ?’ ‘If I'm looking to do X, where would I find Y?’ Those types of questions.”

Is the type of content that's more likely to be surfaced likely to be a certain part of the buyer journey?

“Yes, usually. For example, if you're choosing to buy a software – let's say you're a CFO and you're choosing a new financial software – often, in that buyer journey, part of your process is going to be comparing different solutions.

So, you want to make sure that you are answering questions that might be a CFO comparing different financial software to each other.”

Got you, okay. Do you have any thoughts on the volume/the type of content as well? Because, obviously, you can produce voluminous answers to certain questions that may not necessarily require that degree of complexity in your answer.

How do you know when enough is enough when it comes to producing content, publishing content, and how do you know what type of content is best to write or even publish?

“Our general rule tends to be around 100 queries, so find 100 queries in your customer journey where you can answer that will bring up a recommended solution. That's kind of the first step: Let's figure out 100 things that my customer might ask along that buyer journey. You may have a few different types of customers, which will make that a little bit easier because you can break up questions within your different ICPs.

The next thing you're going to do is you're going to answer those questions for them, but you want to make sure that that content is very information-rich. Unlike SEO, where there were best practices around length of article (the number of times the keyword showed up in that article, etc.), AEO doesn't work like that whatsoever. You can have answers that are two to three sentences long, and that still might be referenced. Length of content is not as important as how credible and information-rich that content is.

What credible means in AEO is: Are you the right company to answer that question? So, are you in the space? Are you in this industry or niche? Are you giving a generic answer, or does it have an opinion? Does it look different than the other answers on the web?

Are you an expert? We see this in law a lot. Lawyers are experts, so it's much easier for them to create information-rich content because they are experts in their particular field of law. AI also needs to deem it credible, which often means that it's consistent with what you're talking about in other places on the web – so, if you are a law expert, you're talking about law, and you're answering a question about law, your other content around the web should also be legal content.

That's how they deem you credible, which is a little different than SEO, which is backlink-driven.”

Okay, so relevance as well as authority. It was interesting that you talked about the company being credible to begin with, and then you actually started talking about individual experts and gave the examples of lawyers within a particular niche sector.

What tends to be more important: building up a company brand that has that authority, has that relevance, and gives confidence to search engines that the brand name is an expert in a particular niche or an individual, perhaps associated with social media, as being an expert in their own right?

“The short answer to that is it depends. If you are a software company, often the company is going to be what's credible. If you are a service business that's based on an individual, having that individual as credibility will matter because that's your point of differentiation. You might be hiring that person.

Let me give you an example of this. We work with an executive coach. Her business has a name, but she's also known: she's an Emmy award winning broadcast journalist. She teaches executives how to speak on camera, so a lot of her business is around working with her and her bringing her expertise. In that case, a lot of the content is around her as an expert and is written by her as the expert.

If you are going back to that example of the financial software, if you're a financial software, it's not as much about a person being the expert rather than the company that's building a tool to allow you to do your job more effectively. You're going to have to make a call depending on the type of business you are and where your credibility and differentiation comes from.”

Can publishing video content also add value when it comes to featuring in AI search engines?

“It can – we just don't see it come up a lot, meaning we don't see YouTube as the most mentioned source very often when we do an analysis on an industry.

That said, for companies that are already doing YouTube and they're doing credible, information-rich content on YouTube that's answering some of these questions, that can accelerate your visibility in the AI search engines.”

It's interesting that you jumped to YouTube immediately with reference to video, and I'm sure that most people do, to be honest with you, but you can obviously publish video using other platforms, maybe Vimeo or Wistia, and actually just house it on your website, so it's associated more strongly with your website.

Is that likely to be more of a benefit?

“One thing that makes YouTube different is that they're transcribing the content.

Gemini is already on Google, and so it has access to the transcription of the video, so it's going to know what content is in that video versus if you're posting a video on your website, for example, the LLMs might not be able to understand what's actually in that video other than the description that you give.

What makes YouTube a little bit different is that everything said in the video is going to be transcribed and fed to the LLMs.”

How do you measure success?

If we're not talking about volume of keywords and perhaps fewer people are visiting our websites and only consuming what they find directly on search engines, how do we know that they're doing that and how do we tie that back to actual business?

“Just like in SEO, the end goal is often a conversion or a lead. Same thing in AI search. The end goal is: Did someone find my business through AI search and choose to give me a call or get in contact or make a purchase? That's always the end goal.

Just like with SEO, traffic, impressions, and all of those are leading indicators to tell you what types of content or things are bringing someone into my website to visit that website that may turn into a conversion. The same thing is the case in AI search.

We look at mentions: How often are you being mentioned. We look at your AI visibility, which is basically your probability of showing up for the questions that you care about versus the other businesses that you're in competition with. We use those as the leading indicators but, ultimately, the goal is to lead to a conversion.

When we work with a business, we are working closely with that business to understand how many of the people who are calling in or making an inquiry said they found you through ChatGPT or through an AI search.”

What AI search engines are the most important ones at the moment?

“We see that most of the traffic is still ChatGPT. When we look at Google Analytics that show actual click throughs to a website, 90% of that is still ChatGPT.”

Is there an issue with selling the need to get this done internally, to stakeholders who don't necessarily understand and appreciate the importance of it?

“We're not encountering that. I think a lot of CMOs are very curious about AI search.

They know the tides are shifting, especially with the updates to Google recently that have led to a lot of drops in traffic. I think it's on a lot of companies' minds. I don't think it's a hard sell, currently.”

I'm talking about being on a company's mind/on people's minds. Obviously, things are changing rapidly here. There are AI search engines that are appearing almost overnight that might suddenly get a lot of traffic.

How does an SEO stay on top of things, and how do they plan ahead? What's likely to be coming down the line in six months’ time, and how do you try to ensure that you're ahead of the curve instead of actually just operating and reacting?

“Ooh, this is a tough one. Just like in the game of SEO, you're kind of at the whim of the algorithm changes. There's some of that in LLMs, but the advice that I always give is that if you're doing good marketing, over time, you will be rewarded.

What I mean by good marketing: you have a clear message, that message is consistent, it's differentiated, and you are creating content along your buyer journey that reaffirms that consistency, that message, and that differentiation. Over time, that will help you continually show up in LLMs, even as the models change.”

Jenna, what's the key takeaway from the tip you shared today?

“The top takeaway I would give is that your website matters a lot and the content that you develop matters.”

Jenna Hannon is Founder and CEO at Hatter. Find out more over at GetHatter.ai.