Think multimodally

Irina says: “Think multimodally.

A consistent trend that we've seen growing over time is the inclusion of different types of media in search results. AI can already parse images and videos really well, and it can also explain what's depicted in detail.

By thinking multimodally, you will increasingly capitalise on this, as AI chunks related concepts across formats into a tailored presentation for the user.”

Does multimodal just include video, audio, images, and text?

“It also includes things like infographics and diagrams, and it’s about the way you think through what kinds of content and visuals to use, like avoiding stock images and filler, and no longer publishing walls of text.

In general, provided that you do things according to best practice and ethics (which includes citing your source), you can definitely use AI-generated content. Google has explicitly said in its best practices that AI-generated content, whether it's visual or written, is okay to use on your site or elsewhere.

That's not an issue, but you need to be properly citing whatever you put out there so that you don't accidentally deceive your audience.”

Should you incorporate different elements as part of one article, or does it depend on what is relevant for the particular piece?

“There's no hard or fast rule, but it's very much about speaking to your unique audience's pain points and needs. I'm always rattling on about this being our true north. It's always about being as helpful as you can with what you put out there.

However, it's always important to enrich blog content with meaningful visuals and videos. That's essential for AI-driven search because AI increasingly interprets and surfaces content across multiple media formats. Even before AI came around, stock images and walls of text were never really helpful for humans, so it's not a new policy or rule here.

Being mindful of what your audience really finds helpful is key. That is the true north. If that's going to mean diagrams, infographics, and explainer videos, then definitely seek to create a diverse portfolio of ways you communicate your story to your audience.”

How much volume is a wall of text?

“At this point, I would be wary of anything that goes beyond 1,000 words.

Again, it depends on what industry you're talking about. If you're talking about beauty and health and wellness, or more visually-focussed markets, that number might dwindle even more. If you're speaking about highly technical markets like the aerospace industry or something of that nature, then those walls of text might be more relevant.

Again, there's no hard or fast rule. It really depends on the ecosystem your brand is in.

There are still places where pillar articles that almost read like a short book are still helpful, and there are certainly audiences that will find that useful. However, nowadays, because of the way AI parses information, it's more about increasing engagement and comprehension. That's why a multi-format way of approaching content is going to be more helpful.

I don't know if those really long-form pillar posts were ever that useful, to be honest. The average reader's attention span is only getting smaller over time. Even back when we were preaching that, it was helpful to consider more ways of communicating to people, and more ways that they might be able to digest that information.

At school, we were often taught that there are different ways of processing and absorbing knowledge. There are visual learners, there are audio learners, and there are learners who are more focussed on taking in text. That's still the case. Not all of us parse info the same way. Thinking about it that way, whether there's AI or not, will help you get to the largest number of people who are interested in what you have to say.

Why do you believe that content should be presented in a modular fashion, so that each section can be pulled independently for an answer box or summary by AI?

“AI would struggle to use content from a pillar article of 5,000 words, even if it was structured optimally.

Modular content is crucial because AI will increasingly extract independent sections for snippets, voice assistance, and multimodal presentations. Each section should have a clear heading, summary, and optional bullets, if possible.

For example, a B2B SaaS white paper can still be split into independent case studies, metrics, and feature benefits. Making it easier for AI to parse and extract what it needs for that particular answer is going to serve you better in the long term.”

How do you know that your content is optimally produced for AI?

“While we don't have nearly as clear a visual on this as we would with traditional SEO, through Google Analytics, there are a lot of little things we can look at to get a better sense of what's going on.

For example, you can religiously track your featured snippet appearances by having a list of keywords that are important to you and regularly checking in to see what your presence looks like there. You can still do the age-old things of looking at bounce rate, conversions, and engagement metrics, which are all available in Google Analytics.

Something else that might be helpful for looking at cross-platform engagement and conversions is looking at the AI-generated answers that are most relevant to what people might search to get to your business. While there are no analytics for that yet, you can track this yourself by just staying on the pulse of what pops up in questions relevant to your own brand's discovery.”

Why should SEOs expect user modelling to expand?

“Google has long been able to personalise search results based on user behaviour and interests, and that now expands to AI platforms. I predict that, in 2026, AI platforms will tailor generated answers to the type of media the user prefers to interact with.

Personalised search is becoming more sophisticated. Media preferences are still more of a prediction, but we're likely to see AI considering behaviour, media preferences, and interests more closely. Users who prefer video over walls of text may see more tutorials served up to them. Users who prefer text may see more detailed guides served up to them. B2C brands can offer really short-form videos, interactive content, and social snippets.

There's a lot of opportunity here to consider the way in which different members of your audience might take in information. It always comes back to this multimodal approach.”

How do you determine the likely impact of a piece of content if you don't know who it’s going to end up in front of?

“That's one of the biggest things we have to think about as SEOs.

It's about taking time as a business leader to figure out the core essence of what you're trying to tell your audience, independent of how you serve it up. First, sit down with your team, or by yourself as a business leader, and tell the story of your brand. Go back to the pain points of your would-be customers, and always come back to helping them on their hero's journey, as a guide.

You, the SEO marketer, are not the hero. It is your customer who is the hero, and you are here to make life a little bit easier for them. Go back to the absolute bare essentials and figure out what pain points you are trying to solve by being genuinely helpful.

Once you've got this nailed down, take these concepts to your marketing communications team and encourage them to spin them up in different formats: video, text, diagrams, infographics, or whatever it might mean. This will largely depend on what space your company is in.

A B2B SaaS audience is used to having content served up in a different way than a beauty and wellness product. Some of those things have not changed, and people still expect certain ways of parsing info from certain places. Alongside these older strategies, however, it will be helpful to have your marketing team take those core ideas you've come up with and create different ways for people to parse them.

Then, just make sure your analytics guys stay on top of analytics for each of them, using whatever is available at this time. There will be more tools rolling out to be able to better track and measure things from AI search.”

If AI likes your content, is it happy taking your content from different places, or should you just publish one unique piece in one place?

“Duplicate content is still very much an issue, and we do not want to flood the internet with billions of iterations of the same thought. That is precisely what Google is trying to clean up.

Common sense would say that AI also does not want to be fed a thousand iterations of the same idea from a thousand different places. If you're going to tell a similar story on multiple platforms or channels, stick to making it as unique, helpful, and tailored as you can.

Maybe there's a core idea there, but make sure that you are always asking yourself whether this is adding anything new to the conversation, that is meaningful and helpful. That is still important, and it’s in everyone's best interest to avoid duplicate content.”

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

“A lot of these ideas are nothing new. People parse and learn in different ways. People hang out in different places on the Internet. Now, with AI coming in, it's more critical to speak to different audiences and learning types.

It’s helpful to think multimodally, create content in different forms and formats, and honour that, but at the same time, stay true to the core story and pain point that your brand is trying to solve.”

Irina Papuc is Co-Founder at Galactic Fed. Find out more over at GalacticFed.com.