Why accessible content is AI-ready content

[00:00:20.500] Will Noble:

Good morning, Jake.

[00:00:22.140] Jake White:

Morning, Will. How are you?

[00:00:24.390] Will Noble:

Not too bad, not too bad, how are you?

[00:00:26.540] Jake White:

Yeah, very well, trying to keep cool, I think, as everyone on the webinar will be as they're joining in.

[00:00:32.250] Will Noble:

I have noticed that we're both dressed like we're gonna go to the Sahara, so…

[00:00:35.810] Jake White:

Yes, that wasn't an intentional ploy, was it, but…

[00:00:41.160] Will Noble:

Straight up to this, we can go into an environment and stay cool.

[00:00:45.580] Jake White:

Ice cream in the Sahara, I think, is the way.

[00:00:48.390] Will Noble:

Yep.

[00:00:51.350] Jake White:

Yeah, good morning, everyone. We'll just give it a little bit of, probably another 30 seconds or so, just to let everyone join, and then we'll run through a few bits. But just while everyone is joining, there are a few bits that we'll be talking about around AI visibility, specifically around accessibility, but a lot of what we're going to be talking about is AI visibility today.

If there is an interest from your side in AI visibility – hopefully there is, because you're joining this webinar – it's a huge thing that everyone needs to worry about at the moment. You'll see on the bottom right of the screen there is a QR code that will give you access to a free AI visibility report that will be run by Squiz for you, and then provided at a later date. What we're looking to do there is give you a bit of a snapshot of where you stand at the moment, a view on where your content is good, where it falls down, and then some actionable steps for you to take away. I'd strongly recommend people go and have a look at that visibility report, if you haven't already.

I think we're kind of at 2 minutes past, so everyone can join. As it says there, there's a recording available after this session. For a quick bit of introductions before the housekeeping – I'm Jake. I'm an account director with Squiz. I've been with Squiz for nearly 10 years, 10 years in a couple of weeks, which is very exciting. I spend a lot of my time working with higher education and local government sectors. And obviously a lot of what we're talking to our clients about at the moment is around AI visibility and accessibility. Will, you're here with us as well.

[00:02:40.220] Will Noble:

Brilliant, thanks, Jake. Hello, everyone, my name's Will, Strategy Director at Squiz. Coming up to my 11th year at Squiz, in various different guises. I've worked with the concept of accessibility for many years, as part of our enterprise search team. And most increasingly now, working for the Digital Experience Platform, with a particular focus on professional services. Very interested to dive into this topic. Should be good.

[00:03:04.100] Jake White:

Yeah, thanks, Will. Just as the housekeeping – we have a recording that will be available post-session, that'll be emailed out pretty quickly afterwards. If there are any questions, please pop them in the Q&A box, not the chat. We're going to leave the chat for any technical issues – that is being monitored. We've got some time set aside for a question and answer at the end, but please do drop questions in as they come to you throughout the webinar.

What we wanted to do first and foremost – hopefully a lot of you know who Squiz is, but if not, we just wanted to give you a bit of framing around Squiz and why we're here to talk to you about this. We've been working in market for beyond 25 years now, and we provide, as Will mentioned, a digital experience platform – a suite of very tightly coupled tools that allow you to create brilliant digital experiences. We have an enterprise search tool called Funnelback, which a lot of you may already use, and we have a Squiz Content Intelligence tool that is live – it went live last week – which is focused on content health and how your content is viewed by AI tools. So that's part of the reason why we're here to talk about that. We're not going to go into the full DXP today. If you want to do that, please do get in touch with us. We'll share our details at the end of this webinar, but we're really going to focus and deep dive into how accessibility can really act on your AI visibility issues.

So this is just a bit of what we'll cover today: framing that problem for you, looking at why accessibility is a really core thing, what your content needs to do, what you need to do as a content team, and then a fix list and how we can actually scale that and keep improving.

[00:05:05.460] Jake White:

So, jumping into that – as Will and I mentioned, we're both working with a number of clients who we're talking to about AI visibility, and specifically accessibility around that. And on screen, we have a few things we've heard from clients over the last few months.

At the beginning of 2025, AI was viewed as a bit of a viral novelty. I'm not sure that everyone knew that it would get to this point in 2026, where it's actually a daily tool that people are using. That's a huge rate of change in the last 12–18 months, and it's a massive change in human behaviour – a massive change in the way that people search for content and view content. And it's really, really hard to keep up with and understand how to react to that rate of change and where to put your priorities.

Actually, just to give you a view of that rate of change – yesterday I was talking to a university who were looking at this as a topic. They actually ran a check to see how ChatGPT referenced them, because they were worried they weren't showing up. One of the main courses that was their speciality – they actually weren't referenced in the area that people were looking for. So if they were looking for business in London, for example, the university wasn't showing up, which was a real problem for them obviously. And there's no way to track how many students they've lost through that. It's a really big thing that people are looking at. I don't know if you're hearing the same sort of stuff in your area, Will.

[00:06:53.840] Will Noble:

Yeah, interesting anecdote, yes. Similar – quite a lot of top-down pressure from executives as well. I was talking to a professional services firm; their partner came to the director of marketing and said, I searched on Google: "who can help me with a commercial contract in the South East?" – and in the AI summary, this firm wasn't cited. And this partner asked: what are we going to do about it? That's a big question. To be able to understand and answer that confidently straight away is tricky. But we're also seeing some interesting shoots of opportunity at the same breath.

I worked with a firm recently that just won a six-figure engagement from a lead that came through GPT – this prospect was running all of their informational searching, understanding about a practice area. This firm was cited, and when they came to the site and did that inbound, there was a lot higher level of intent, so it actually closed a lot faster than usual engagements. Yeah, a lot of worry and uncertainty, but at the same time some interesting opportunity.

[00:07:56.900] Jake White:

Yeah, huge opportunities. And part of that conversation I had was also people not knowing where to start. It's a big thing – we know that's the problem, we know that you can't track it, and we know that we need to really author our content to address that. I think you brought up, when I had a conversation with you about it, Google's plans to sunset the 10 blue links as well – which is a show of how big this is becoming. What we really want to look at now is how do we solve that problem, and framing that problem as well.

[00:08:32.400] Will Noble:

Yeah, good setup there, Jake. So, since 2022, we've all faced the same issues. We all have AI visibility problems. Squiz has a visibility problem; the clients we serve have this issue. We're all seeing traffic decline – around 30%. And that is driven not just by the advent of the large language models we know of, but also the big layout change that's happened on Google. Google I/O last week said that they're going to get more interested in answer engines and conversational experiences, rather than 10 blue links. And even in that blog Google shared, they've seen a 30% increase in brainstorming queries. So: "what should I do here?" Typically we used to search on Google with just a couple of keywords – that's not the case anymore. But this traffic decline results in a bit of a brand problem. You've got fewer individuals seeing your brand, understanding what makes you distinctive, understanding where you fit in the market. And if fewer people know of your brand, that manifests itself in the long term into a revenue problem. So, how we react as marketers and digital professionals to this new paradigm is the interesting bit.

In previous webinars, we've talked about this AI visibility problem from the lens of content and content structure – you can find those on YouTube. But understanding more about how we can be cited in answer engines is definitely a strategic advantage. And the good news is, the work that you may already be doing on accessibility turns out to be exactly the same work that prepares your content for AI. It's something most people haven't joined up yet, and that's what we're going to look through in this webinar.

[00:10:27.570] Will Noble:

So, accessibility is AI readiness. While researching this topic, I went to Gemini, and Gemini coined the term that AI is the world's most powerful blind user. Which encapsulates the complete hubris that these models have. But the machine does kind of have a point to some degree. A sighted human looks at a web page and instantly understands what's there through visual cues – like font sizes, spatial layouts, sidebars. AI crawlers and screen readers don't do this. AI doesn't see the visual layout in the traditional sense – it sees the underlying document structure. Both an AI crawler and a screen reader parse the code, and they rely on the explicit structural cues to understand what's important and how all of these ideas connect.

So, if your website code is messy, then an AI might misunderstand the context of your content and overlook you for citation. We're going to go through some of those structural cues in a second.

I guess there are two wins – actually, it's probably three wins – in making sure your site's successful. The first thing is it's just the right thing to do. 25% of the population, especially in the UK, have some form of either temporary or permanent disability. So excluding that audience is bad for your business. In the UK, businesses lose an estimated $11 billion annually because customers abandon inaccessible websites – and that's the same if you're servicing citizens, students, or B2B professionals.

The second win is that you're working towards compliance because of regulation. We've got the EU directive, ADA in the US – public sector orgs have accessibility targets, especially if you've got a big, sprawling digital estate. So the accessibility work you're doing is even more valuable. And then the third win is this paradigm we're in now, where clicks are less important to your brand. We're hunting for citations, so working on making your site accessible increases your chances of being read and cited by large language models.

[00:13:03.700] Jake White:

I like how you switched from 2 to 3 there! But yeah – the good news is that this is something that people have been working on for years. Accessibility is not a new thing. WCAG guidelines are not a new thing. Certainly in the space that I work in – higher education and local government – it's a huge thing that people have been talking about for as long as I've been here. And as you say, it's not all new stuff. With GDS guidelines and WCAG guidelines, everyone's trying to adhere to those based on the area they work in. And actually – as you said – accessible content is visible content.

[00:13:54.080] Will Noble:

So we've pulled out some of the common fixes in the middle of this slide. On the left, you've got why screen readers need it, and on the right, why AI needs it.

AI agents view websites in three ways. They can take screenshots, but that can be resource-heavy and expensive. Or they look at the raw HTML, or they use the accessibility tree. An accessibility tree is basically a way of the browser giving a semantic summary of the page – which is the same thing screen readers use. So there's a huge overlap here. When you're optimising for WCAG, you're inadvertently doing answer engine optimisation, which is really valuable.

Should we pick out a few of these and go through them? The first one: heading hierarchy. Headings establish the hierarchy of ideas. H1 is your main topic, H2 is your subtopic. These headings make it easier for a visitor using a screen reader to get a summary of the page – but AI uses this to map the relationships between concepts. Two birds, one stone.

And that ties into semantic HTML. Tags like header, article, or nav allow you to skip to the main content and navigate effectively if you're using a screen reader. For an AI crawler, this is basically a content filter that provides context to the page. If you just use divs across an entire page, it's hard to understand the context of what that copy is.

Some of the newer things at the bottom – about vague copy and having explicit language – this is something that Squiz's Content Intelligence tool does quite well. We ran it across the Squiz website recently. The phrase "enterprise-level security" is cited a lot. It's great, you can kind of understand what that phrase means, but if you think about it more deeply, what does that actually mean? Enterprise could mean something different to different organisations. Instead, you need to say: we offer enterprise-level security to ISO 27001. Give the accreditation. Really small structural changes to the way that we describe things have a big impact on citations in LLMs, but also help people who have accessibility issues.

[00:16:38.160] Jake White:

Yeah, and I have an example of that from my own experience. My wife – she's Australian – was going through the process of trying to get her citizenship. Obviously, like everyone does now, I went to Gemini or ChatGPT to go, okay, what are the next steps? And the amount of legislative language that came out – there was absolutely no way to go, okay, this is the next step. If you have a clear-sighted breakdown: step one is this, step two is this, step three is this – that becomes much easier for LLMs to read. They're not going to spend loads of time and resource trying to find content for you. They'll find whatever's as easy as possible to push and present to the user. That's why that one is really important.

And one of the things on here that always jumps out at me – it's quite a joke within the team – is around keeping structured content in PDFs. Local authorities are notorious for lumping meeting minutes and everything into a PDF and uploading it to the webpage, then having citizens make complaints about not being able to find the right information, or sending people to the call centre because they can't find that information – which is an incredibly expensive way to deliver that answer. So another real bugbear for me is making sure we put really key information into structured pages, in that specific, explicit language, which has that dual benefit of being accessible and also AI-ready.

[00:19:03.500] Jake White:

But for everyone here, we don't just want it to be Will and I talking for 40 minutes. We just wanted to run a bit of a poll to keep it interactive – looking at: when it comes to your content's accessibility and AI readiness, do you know where to start?

[00:19:50.220] Will Noble:

Should we give everyone a minute to do that? While we wait, Jake – if I was to say the number 3.4 billion, what would that speak to you?

[00:19:41.260] Jake White:

Initially – a rather large number, is what I would come back with without knowing the context.

[00:19:50.220] Will Noble:

So, $3.4 billion is the AI budget that Uber burned through in 4 months using cloud code. That's their entire yearly budget gone in 4 months.

[00:20:07.230] Jake White:

Wow. No wonder people are talking about putting servers on the moon.

[00:20:27.890] Jake White:

Cool! While we hopefully have the answers in the poll – I'd be interested to see what people have said, Will, if you wouldn't mind reading a few out.

[00:20:42.770] Will Noble:

Yeah, sure. So, the question is: when it comes to accessibility and AI readiness, do you know where to start? 19% said yes, we've got a clear priority list, which is really good. The majority – 37% – said we know there are issues, but aren't sure where to fix first. 24% said we don't know how big the problem is yet. And a smaller cohort, 20%, said we know what needs fixing, but don't have the resources to act.

[00:21:10.970] Will Noble:

That's actually a really broad mix. I thought there'd be a couple of answers that took the most. Most of them realise they've got issues, but it's just: how do you take a bite of the elephant? Where do you focus your effort first?

[00:21:28.840] Jake White:

And that's a big thing – how do you eat the elephant? It's bite by bite. We're going to look at how we can pull that together. I know you've got a rather natty Venn diagram that you wanted to run everyone through.

[00:21:50.340] Will Noble:

Yeah, my pride and joy. So, we're building sites for three audiences now: for humans, for accessibility, we've been building sites for search engines for two decades, and now in the last couple of years, we're building sites for answer engines. But AI, accessibility, and SEO don't work in isolation. They all reinforce one another. When we were putting this Venn diagram together, we were actually struggling to pick things that are distinct and don't interlink – because it's all related.

A lot of these things people will be familiar with, like having good alt text, clear metadata, and JSON-LD. But in this new world, we're hunting for citations, not just clicks. There are a few new things too.

Traditional SEO always encouraged long-form, comprehensive articles, and now it's all about writing content in chunks. Every section under a heading must now function as a standalone paragraph. Which I actually think is a good thing – you know the classic story of trying to find a recipe on those websites with the huge life story before you get to the actual recipe, just padding keywords into the article? Hopefully that's going to go away.

Content also now has to optimise for something called query fan-out. When a user types a prompt into an answer engine, that answer engine breaks the question down into lots of component parts and runs additional questions behind the scenes to formulate a response. Which means you may not rank for that initial keyword – which used to be everything in the SEO world – but you can be cited in an answer for those secondary questions. It's actually quite democratic, and it does mean that your content on your website is going to expand.

One final thing on this Venn diagram: SEO for the last few years hasn't really been affected by JavaScript-heavy websites. You could run PWAs and have everything in JS, and modern crawlers would grab it and still rank it. But modern AI crawlers built by different providers don't execute complex JavaScript the way traditional Google or Bing does. They'd much rather just extract the raw HTML. And of course, raw HTML is great for screen readers and accessibility. So all this is to say: don't throw out your existing playbooks, we just need to add to them.

[00:24:52.110] Jake White:

Yeah, and I think, touching on that as well – hopefully everyone on this webinar has been focused on accessibility in some form. There's always a fix list of things to go through, and hopefully nothing on the screen is overly scary or new – everyone knows roughly what they need to look at.

But there are a few bits in there that are incredibly important. We've got my two bugbears already – moving content out of PDFs, and removing vague language. I will die on that hill. But we know that everyone's been going over content fixes for years: alt text on images, hierarchy, things like that. There are things that don't get brought up as often, like site speed and semantic HTML, as you mentioned earlier. This slide is really just to call out that what you're doing at the moment as a content team is brilliant, in terms of keeping that up – keep moving things out of PDFs.

Also, if we're trying to remove things from behind a login wall – don't put really important content behind a login wall. That can lead to real issues in terms of the way that AI is viewing it, but also from an accessibility standpoint as well. Site speed is obviously really key. We want to make sure that everything loads quickly. Slow-loading sites are not only inaccessible – they're also skipped by AI crawlers, and we'll touch on that in a bit.

[00:26:45.560] Jake White:

So, into the more meaty stuff. We've done the setup, now we want to look at how we can actually really focus on this. What we really want to call out is: this needs to be a continuous cycle. Gone are the days of doing a mammoth accessibility audit at the start of the year, working through it for 12 months, and kind of sticking it on the back burner. We really need to take those small wins, audit, prioritise, fix, track, and repeat – rinse and repeat all the way through.

As mentioned – don't try to boil the ocean. Take the really key parts of the site, the really key focuses for your organisation, and make sure those are the top things you're looking at first.

So, as I mentioned, a lot of people I've been talking to are going through those mammoth yearly accessibility audits that generate a spreadsheet, and everyone goes: oh god, how do we deal with this? I was actually talking to another university a couple of weeks ago who have two full-time members of staff just going through spreadsheets, putting in manual fixes for accessibility auditing. They've actually just got to a point where they've built their own AI agents to do that – but both of those approaches bring a huge amount of human error. And it's pretty overwhelming. If you don't have the area of expertise, you can cause more damage than good. So what we want to do is make sure that you work across the whole site, but chunk it up into small priorities, and not end up with something that makes you feel sick every time you look at it.

[00:28:59.490] Jake White:

So, where that comes in – we have the old-school accessibility audits looking at spreadsheets and having all of that. But in the new world of AI discovery, it's making organisations really look at whether their current accessibility tools are working for them, and whether it's time to consider the value they're getting from those tools. Whether they have the right capabilities moving forward, because most tools are mainly built for the traditional accessibility fixes: heading structure, colour contrast, alt text, compliance to AA, and getting those static recommendations.

Whereas what we want to start looking at is also: whether we've got vague copy, conflicting information, questions that can't be answered from your content, and giving you an AI readiness score. The two are symbiotic. If we just have the traditional accessibility side, you'll do fine, but you won't be as optimised. We really want to make sure that you're pushing your accessibility providers and AI readiness tools to make sure you can have the emphasis on both sides of the coin, rather than just one.

[00:30:41.560] Jake White:

So on that note, we just wanted to ask another poll: are you using your accessibility tools to their full potential? We want to know whether you're very confident about that; whether you have some tools but aren't sure about the value you're getting from them; whether you rely on a super user; or whether you don't have an accessibility tool at all.

[00:31:34.690] Will Noble:

Do you want another number question? What does 2013 mean to you?

[00:31:43.730] Jake White:

Oh, 2013. Trying to think of something that would have happened – good Ashes series, I think, 2013.

[00:31:51.980] Will Noble:

Okay, it's topical. It was the year that we open-sourced Squiz Code Sniffer. That is still open-sourced – it's a browser plugin, and it just goes page by page, which still has a lot of value. That's when we first started building some of this technology for this concept.

[00:32:17.160] Jake White:

That's why I love doing these things with you, Will, just to get those little snippets. So yeah, let's have a quick look at how that poll's come out.

[00:32:36.200] Will Noble:

So, the overwhelming majority – 50% – said we have tools, but we're not getting full value from them. 25% are very confident. A smaller percentage, 12%, said we rely on one super user and it doesn't scale. And 13% don't have dedicated accessibility tools.

Knowing the audience that we do, I think we've got a good number in the public sector and universities, federally mandated to get this right, so there is a good proportion of you getting your value from these tools – which is great, good advocacy for accessibility. And even though 2013 is when we first started building some of this technology, having 13% that don't have tech on this shows that there is still a long tail for looking at this concept, and these webinars are still valuable for a good cohort of attendees.

[00:33:40.780] Jake White:

Yeah, and I think what we're going to do now is just jump into what's possible when you do have the right tool working for you as well – I'll hand over to you, Will.

[00:33:53.210] Will Noble:

Yeah. So, we had a few different iterations of this, and our enterprise search engine shipped with an accessibility auditor to measure conformance over time – which is a great way of sole-sourcing our tech. But because of this connection between AI visibility and accessibility, we've now reimagined a platform to enable you to measure conformance over time, but also prioritise issues in terms of their severity.

In this new and improved experience, rather than just telling you you've got an issue, we actually give you remediation advice – here is a snippet that you could use. What I quite like about this is the way that it isn't just focused on the developer in your team, because a lot of organisations don't have a developer in-house. You may outsource that to an agency, but we give you advice that could be valuable for a content user as well – about the words on the page. And it's then tied into all of your content health as well.

That came out last week. If you're an existing client, you can speak to your account manager to get a view. If you're not an existing client, you can scan that QR code.

[00:35:09.020] Jake White:

Yeah, and one of the really cool things that the Content Intelligence tool does – what people struggle with is how do you prioritise that part? We can actually break that down into topics as well, and specific parts of the site to look into. So it's not necessarily just looking at the worst parts of your site – it's looking at the bits that are going to provide you the most value. And it does that from an accessibility standpoint as well, which I think is really fantastic.

[00:36:06.590] Will Noble:

I can capture this one. Jake talked about it at the top of the call, but if you're interested in an AI visibility report – which goes into depth against your own content for where you can prioritise – we're offering those as a free take-home for today. Please let us know if you'd like one.

Let me do a quick recap before we jump to questions. What have we gone through? AI crawlers have changed where decisions are made – we're not chasing clicks, we're chasing citations. If your content and page structure is inaccessible to these crawlers, you're not going to be cited. The good news is that working on accessible sites helps with answer engine optimisation, so it's really valuable and a good mandate to continue working on this subject area.

Our recommendation is to take a look at your existing tech stack. Is your accessibility tool providing enough remediation advice, and giving you enough indication about where to start? If it is, brilliant. If not, it may be time to take a look. And as we talked about, Squiz Content Intelligence can automate the audit for you and give you some recommendations about where to start.

[00:37:27.290] Will Noble:

Right, let's do some questions. Where to start is actually our first question: how can we identify which pages to prioritise first for AI and accessibility improvements?

[00:37:41.920] Jake White:

I think – as I sort of touched upon – it's dependent on the priority of the organisation first. One common misconception is to pick the worst performing parts of the site first. Although that is a valid approach, we want to focus on the bits that are providing you the most value as a business. Whether that's looking at courses if you're a university, or voting pages during a general election if you're in government – those are the bits I would encourage people to look at first. And then go from there into the lower-priority pieces. I think the structures you put in place on the priority pages will naturally flow down into those. I don't know what you think, Will.

[00:38:46.730] Will Noble:

No, I would agree. Where are your most high-traffic pages? Which ones drive the most value for your business? If it's a legal services firm, what are your service pages or your most prominent lawyers? Because if those are the areas you want to be cited most, you need to make sure they are up to scratch. And often, from a university perspective, your worst performing pages are things like physics.university.ac.uk that haven't been updated for 14 years and are on an old platform that's very hard to change the markup on. You can spend a lot of effort trying to engage with stakeholders who own that domain, rather than focusing on your primary www. Really valuable.

[00:39:37.070] Will Noble:

Next question – do we have any suggestions for accessibility tooling?

[00:39:41.920] Jake White:

I would suggest that the Squiz Content Intelligence tool is a fantastic accessibility tool. It ships with that. Our product team are very heavily invested in improving it as well, and we're able to get client feedback onto the roadmap really quickly. We really engage with our clients to make sure any improvements they want to make are dealt with very quickly. So absolutely – if you're looking for an accessibility tool, please do get in touch with Will or myself, or if you're an existing client, with your specific account manager. Please do start using the Content Intelligence tool in earnest and provide your feedback.

[00:40:55.150] Will Noble:

There's a question here from Innes. They're in the process of implementing AI conversational research – it's been a huge learning curve. They've heard a lot about schema markup. Is this a priority if their accessibility and content is already in a good place?

We see this a lot on sites. We have anecdotal evidence that it enhances your visibility – these AI engines love structured data. If you've got JSON-LD on some of your pages, it can sometimes help with citations by around 20% using brand-mentioned tooling for some of our clients. So if you have the bandwidth, it's certainly something I recommend looking at and adding to your headers.

[00:41:47.140] Jake White:

Excellent. We've still got probably time for one more.

[00:41:53.340] Will Noble:

There's an anonymous one about web content: should web content be formatted in a way that mimics answer engine outputs – basically Q&A formats and lists?

I think the answer is, there's a toss-up. There are some pages where there's going to be a chunk of 4 or 5 paragraphs that want to be related to each other. If we write in a way that's just Q&A, it feels a bit disjointed for someone actually on that page. Our approach is: yes, putting frequently asked questions onto your website is a good idea, tied directly to that key page. A separate FAQ section isn't as valuable as it once was – it's usually a bit of an afterthought, where you put in loads of FAQs that no one actually asks.

There are kind of five ways that AI likes content to be: complete, concise, fresh, specific, and accurate. Making sure your copy is written in that way is very valuable. I'll share the deck from the previous webinar where we went through all of those. I think the challenge comes when you're publishing thought leadership that's really 7 or 8 pages long – it's hard to do those connections. So it's knowing what's right for your organisation, and taking those best practices and following that approach.

[00:44:01.010] Jake White:

There's also a common misconception that AI or LLMs understand your content in a Q&A format. Actually, the way it works is that it predicts what the next point will be. So if you're saying "what accommodation is available?", it's not reading it and going: I will only answer something that's in a Q&A format. It's actually making sure your content holds that bit of information in a really concise way – for example: our accommodation is this. It's predicting that and finding that information, rather than understanding the question itself. So just make sure you follow those steps to keep writing your content – as Will said – in your block content and key content in the same way, but making sure it's concise and doesn't have those long-winded answers. Kind of like – if you're going to a party and you ask where the bathroom is, and someone starts going on a long-winded way around the house, or someone just says: it's down the hall to the right. You're hoping for the second one.

[00:45:05.040] Jake White:

One last thing – I know we're nearly at time, but David Higley asked: will the tool work on non-Squiz sites? And the answer is absolutely yes. So if there are clients on here that are not existing Squiz users, or only use Funnelback but don't use the DXP to host your sites – for example, David's was around having sites on WordPress – you absolutely can audit those using the Content Intelligence tool as well. I'll speak to Kia who I think is your account manager, David, and we'll get in touch with you about that one straight away.

[00:46:07.180] Jake White:

Will, is there anything else you wanted to add, or any other questions you wanted to go through quickly?

[00:46:12.660] Will Noble:

I think we're at time. We're a minute over, so – thank you everyone for all your questions. Any questions we haven't got to, we are going to send out answers afterwards, and we'll build a landing page to cover them. The recording of this video and previous webinars that talk about AI search visibility will be on YouTube, on the Squiz official channel.

[00:46:34.260] Jake White:

Thanks very much.