Mel: Remember how we used to spend a lot of time and money on search engine optimisation? Asking other organisations for backlinks to our websites, drafting long form content like blogs, using the right keywords and hashtags, and maybe you're still spending a lot of time and money on these things. But the reality is they may be delivering far less value than they used to,

And the reason is AI. More and more people are now using AI tools like ChatGPT, Claude, and Gemini as search tools. And unlike traditional search engines like Google, brands don't show up in the results in these tools. You also have no direct control over whether or not your company, your program, your product, or your service, whatever it is, shows up above any other competitor.

You can't pay for that, at least not right now. So we're entering a new era where we have to think and write [00:01:00] differently if we want our brand to show up in the best light in AI search results. How do we do that? Well, that's what today's episode is all about. Hello friend, and welcome back to another episode of Less Chatter, More Matter, the communications podcast.

I'm your host, Mel Loy, and I am recording this on the lands of the Yuggera and Turrbal people here in Meanjin, also known as Brisbane. And I've just returned from a trip to Canada where I attended the IABC World Conference, and a hot topic was, of course, artificial intelligence, AI. Whether that was around how we use it, why we use it, the risks in using it, or anything in between, almost every speaker had something to say about it.

And what was clear is that many people haven't yet made the shift to thinking about AI tools like Gemini or ChatGPT as stakeholders who need specific messaging like any other audience [00:02:00] segment. Now, the stats do vary a bit here, but ChatGPT is reported to process over two billion queries every day. Some sources say that around forty percent of people use AI-powered search engines above traditional search tools, and that does continue to grow.

So if you are not d- adapting your content to give it the best chance of success of being surfaced by an LLM or a large language model, then you are already behind. We have entered a new era where SEO is being replaced by GEO, or Generative Engine Optimisation. It's sometimes also referred to as AEO, or Answer Engine Optimisation.

Now, GEO is different from SEO in two key, key ways. The first is that AI tools will search for passages of text rather than for pages like a Google search does. It's looking for a clean, quotable passage of text, even if that page [00:03:00] that it's on is not ranking particularly highly. The second way it differs is that visibility is driven by frequency in terms of how often your brand or content appears in different responses to different prompts.

So what do you need to do to alter your content to make it more GEOable? Let's get into it. Number one: start with the answer, not the question. So think about the questions a customer or stakeholder might be asking, whether it's about your business, the sector you're in, or something specific like a service or a product or instructions, and then provide the answer.

What this means practically is that you should start each section of text with a direct answer before you provide context. You want to make sure your key point is at the top and it's clear and concise. And this is because, like I said, AI loves to pull out the most direct and quotable passages of text [00:04:00] and also assumes that people don't scroll through long content, which look, it's probably right in that assumption.

So for example, let's say you are talking about a new product or service. Now, traditionally, we may have written a media release or website article about this by starting with a bit of context or background on where this new thing has come from before getting to the key and specific point. We essentially have followed that narrative arc of context, credibility, and then the new point.

Now we need to flip it. We need to write with that answer first structure. So as an example, let's say you might be talking about a new chatbot that you're introducing to speed up customer inquiry times. Now, if you wrote in the traditional way, it might sound like this: "Our company has been providing exemplary customer service for more than twenty years. But in this time, technology and expectations have shifted, and people are busier than ever. Our customers have told [00:05:00] us that wait times for our call centres and response times for emails are far too slow. So we're introducing a new chatbot which will help give customers the answers they need in as little as twenty seconds."

Now, if we are taking an answer first approach, then we'd actually start this passage this way "Our new customer inquiry chatbot can solve your problem in twenty seconds." Then you would get into the complex, so why you've done it, et cetera. In a way, it's actually quite similar to how you're taught to write for media, if you've ever done journalism.

It's where we are always starting with the lead, so don't bury the lead. So that's tip number one, start with the answer, not the question. Tip number two, while your structure matters more than ever, so structure has always mattered for SEO, especially for web content. Search engines like Google scan for the H1, H2 headings in your website content, and they prioritise that.

It's one of the reasons why at Cuttlefish we [00:06:00] always say to our clients that they should use the built-in heading styles in their intranet and website CMSs rather than just selecting text and bolding it, for example. Also really good for accessibility, FYI. But from a GEO perspective, LLMs really love these clear heading hierarchies too, but they love it even more if there is just one topic per section.

And if you really want LLMs to froth over your text, then make those headlines very specific. Again, it's like answering a question straight up. The headline should include the details of what you're talking about. So rather than a headline that might say something like, "XX company launches new customer service chatbot," you would say, "XX company launches new AI-powered customer chatbot to solve queries more efficiently than ever before."

It's much more specific. Subheadings should also use specific stats because it provides a very concise, pullable [00:07:00] piece of text. Now, as an example, think about that chatbot example we just talked about. So your subheading could be, "Chatbot reduces customer search query time by up to eighty percent compared to traditional customer service channels like phone calls or emails."

Yes, it sounds a little bit longer, goes against a little bit of what we say around concise writing, but when the audience is an LLM, this is what we have to do. Headings, so other headings in your text should also mirror real questions. Remember, we are thinking about that person who's typing a question into a, a tool like Claude or Gemini.

So for example, instead of using a heading like 'Product Features', use a heading like, 'What does XX product do?' Or, 'What are the benefits of XX products for families?' Or whoever it is, and then provide the answer under the subheading. So remember that we want to mirror those questions that those people are likely to be asking, [00:08:00] and you want to deliver the answer on a silver platter. That's number two, structure.

Number three, cite sources, not just facts. This one is more important anyway in an era of fake news and misinformation. But LLMs also love it when you explicitly state the source of the statistic you're including within the text. So for example, saying, "According to McKinsey's 2026 Workforce Report," blah, blah, blah, is much more GEO friendly than simply saying a fact like, "More people are looking for new jobs this year."

Again, LLMs just want to be able to lift and shift content and regurgitate it to you, so you have to make it easy for them to do that. A bonus here too, if you've ever struggled to get people in your organisation to stop using fluffy words and jargon, guess what? LLMs hate it too. Phrases like [00:09:00] game changing or revolutionary are ignored because they can't be fact-checked, and everyone uses them.

They are fluff. They're just not specific or verifiable. So you can now say to those people in your organisation who like these silly words that AI doesn't like them either. It's not just you. Okay, tip number four, demonstrate your expertise. Now, credibility still matters, and again, maybe more than ever in a world where we're dealing with deepfakes and misinformation farms and so on.

And AI is getting better at searching for what it perceives to be credible information from credible sources and filtering out some of the crap. Now, what does this mean for your writing? Firstly, when you are drafting long form content like media releases, blog posts, website articles, et cetera, always include a byline of the author or authors with their relevant credentials.

Also, make sure you are sharing case studies, your [00:10:00] own experiences, as well as any original research you might have undertaken. Secondly, credibility is also built by publishing on platforms or domains that have authority. So this could be a news site, it could be a site like Forbes, or even just your own website if your brand has good standing and is well established.

Earned media coverage for reputable organisations is also highly prioritised. And thirdly, consistency is key. This is where making sure that the way you talk about your brand and products and services is always consistent. The tone of voice you use needs to be consistent as well, and so on. Because from an LLM perspective, that consistent online presence is a sign of credible content. Another thing to note here, especially if you write media releases, is that your boilerplate message or the About Us section that we typically put at the end of a release should be specific as well.[00:11:00]

So that means including things like when your company was founded, how many team members you have, how many customers, um, where you're located, et cetera. This helps an LLM to accurately represent your organisation when someone asks about it, which is important, or it'll pull from somewhere else, and you don't want to risk that.

You don't want to risk them pulling from another organisation or a random person on Reddit talking about your organisation. This is why it's really important also that your About Us section on your website and your social media profiles is also specific. Okay, so that's-- our fourth tip is about demonstrating your expertise.

Number five, well, it's about keeping it fresh. And guess what? Generating regular content isn't just good for SEO, it's now very good for GEO. Typically speaking, content that is less than three months old is more likely to be cited. So keep pumping out those press releases, LinkedIn [00:12:00] articles, blogs, podcasts, et cetera.

AI is also looking for recent and relevant data. There's a lot of old stats out there. I mean, the internet's been around a really long time now. And as we know, things change quickly. LLMs also know this, so data from the last couple of years is going to be prioritised. It means that as comms professionals, we have to keep staying on top of new trends, new research, and doing our own research to keep those statistics fresh. And tip number six, use LinkedIn more.

So LinkedIn has always been highly indexed in Google searches, and now it's being highly indexed in LLM search results as well. And because the content on LinkedIn is updated so regularly every day, I mean, think of all the users using and promoting and sharing content, it feeds into that fresh content and relevant data that LLMs love that we just talked about.

Now, some of the info here is a bit mixed, and it does change [00:13:00] a bit, but a couple of studies earlier this year found that LinkedIn content has become one of the most cited sources across the major AI tools like ChatGPT, et cetera. In November last year, LinkedIn's domain rank on ChatGPT was in eleventh place.

By February, so just within three months, it was in fifth place. Not only that, but on average, 11% of all AI responses now reference a LinkedIn URL. So that puts it above things like YouTube, Wikipedia, Reddit, and news sites. That's pretty powerful. Now interestingly, articles of up to 2,000 words get the most citations, and posts of fewer than 300 words are also favoured.

So when you are looking at your channels matrix and your channels guidelines, keep those numbers in mind. Importantly, though, some LLMs favour content from individuals rather than company pages. [00:14:00] The company page content still matters. Don't get me wrong. It's still looking for that. It's still highly indexing it.

But if you have been toying with a thought leadership strategy for your business, then you should stop toying and start doing. This means getting senior leaders to share news to their own profiles more regularly as well as sharing it via company pages. And guess what? I've got a bonus tip for you. That bonus tip is use AI to help, especially when you're getting used to this new way of working.

Why not use AI to help you? So, for example, let's say you have a draft blog post or media release. Chuck it into AI, assuming that it's not confidential, et cetera, and ask it to rewrite it or give you suggestions for how to optimise it for LLM search results. All right. We've covered a lot there, so let's go and do our episode recap.

So today's episode was all about why and how you should be writing for AI, and I shared six [00:15:00] tips on how to do that. The first one is to make sure that you are answering a question and start with the answer, not the question when you are writing your content. The second one was your structure matters more than ever.

So make sure you're using those heading styles correctly, but also make sure that your headlines and subheadings, et cetera, are specific. Or maybe even ask the question that somebody else would ask. The third one was cite sources, not just facts. Remember that LLM tools are looking for signs of credibility, and by including the source in your text, you will help to boost that credibility.

Number four, while we're still on credibility, is demonstrate your expertise. So include bylines of your authors, their credentials. Make sure that you are publishing on authoritative platforms and be consistent in the way you talk about your brand. Tip number five was keep it fresh. So make sure you are constantly creating new content, [00:16:00] as long as it's good content, because LLMs typically like content that is less than three months old.

And tip number six was use LinkedIn more because it is highly indexed in these LLM search results, and it's a good way to get your senior leaders to maybe finally get into that thought leadership strategy you've been wanting to do for years. All right, team. That is all we have time for today. As always, if you have any comments, feedback, ideas, questions, get in touch and also, if you haven't yet, I would love it if you could rate and review the show.

I also want to take a quick moment here to thank every one of you who has listened t- to the last few years of Less Chatter, More Matter. We've now reached more than twenty thousand downloads from across the globe, and I could not be more grateful for you listening in. Right. That's it for me. In the meantime, keep doing amazing things and bye for now.