Skip to content
Viewing: all (33)
  1. Alex Capes Your content now travels in pieces SEOContent strategyDigital strategyMarketing
  2. Barry Cumberlidge Out-evidence, don’t out-publish SEOContent strategyDigital strategyMarketing
  3. Michaela Brown Ten Years in Tech, Seven BTFs In, and Staying Curious Digital strategyTechnical validation
  4. Barry Cumberlidge The question is the new keyword SEOContent strategyDigital strategyMarketing
  5. Barry Cumberlidge A contradiction costs more than a gap SEOContent strategyDigital strategyMarketing
  6. David Lowbridge The case for connected knowledge Brand automationContent strategyDigital strategy
  7. Barry Cumberlidge Nobody owns the topic yet SEOContent strategyDigital strategyMarketing
  8. Roger Whittlesea What does collaboration look like at the proposal stage? Digital strategyTechnical validationBrand automation
  9. Matthew Froggatt Internal work has no client Digital strategyMarketingContent strategy
  10. Barry Cumberlidge The gift of time Digital strategyBrand automationMarketing
  11. David Lowbridge From fixed values to behaviours Brand automationUX & UIDigital strategy
  12. Mike Leonard The web is fun Motion designUX & UITechnical validation
  13. Jara Santamaria Martinez Every year is a new brief PlatformsContent strategyTechnical validation
  14. Kristina Radonjic The work between the work Digital strategyMarketing
  15. Guanglun Wu Choosing the right tech stack Technical validationPlatformsDigital strategy
  16. Barry Cumberlidge Security deep dive Technical validationContent strategyDigital strategy
  17. David Lowbridge Startups are prototypes Digital strategyBrand automation
  18. Tara Heal Shapeshifting the Future UX & UIMarketing
  19. Tom Small The theme is the guardrail Brand automationPlatforms
  20. Barry Cumberlidge Your biggest asset in an RFP, your website SEOContent strategyDigital strategyMarketing
  21. Michael Gunner Cache Components: the right content at the right time PlatformsTechnical validationContent strategy
  22. David Lowbridge From static frames to a creative operating system Brand automationUX & UI
  23. Tom Small We build by leaving things out PlatformsDigital strategyUX & UI
  24. Michael Gunner Standardise the foundations, not the brand experience PlatformsBrand automation
  25. Charlie Clarke Letting type move Motion designUX & UI
  26. Andy Purbrick Putting video back where it belongs PlatformsTechnical validation
  27. Mike Leonard The cache that doubled as a coordinator PlatformsTechnical validation
  28. Andy Purbrick We test the revenue path Technical validationDigital strategy
  29. Mike Leonard Two months in Three.js, we started over Motion designTechnical validation
  30. Rhona Mackay The form we actually trust PlatformsTechnical validation
  31. Andy Purbrick The testbed nobody will see PlatformsTechnical validation
  32. Andy Purbrick Twenty-four modules, one wrapper PlatformsContent strategy
  33. Andy Purbrick We haven’t missed a Thursday in 33 weeks Content strategyMarketing

Your content now travels in pieces

Alex Capes 5 min read SEO Content strategy Digital strategy Marketing

Buyers now use AI search to ask with a photo or their voice, getting answers stitched together from your web pages, images and videos.

Last week we argued that evidence only you hold is what gets an AI to cite you. This week is about the form that evidence takes.

Every piece of content on your site now has to make sense on its own, because buyers are more likely to see it as a quoted line or a video clip than as part of a full web page. Your audience is still a person with a question, only now they are reaching you through photos, voice and video as often as typed text. The core work of content strategy hasn’t changed: you still need to know who the information is for, how it’s organised and what makes it worth their attention. What has changed is that you no longer control the order, or the surrounding page, in which a buyer sees them.

Screenshot of a Google AI Mode search result combining a text summary, product image carousel, and YouTube links
Screenshot of a Google AI Mode search result combining a text summary, product image carousel, and YouTube links

How questions and answers have changed

Search behaviour has shifted from short text queries to long, conversational and multimodal prompts. In May, Google announced that AI Mode, its conversational search product, had passed a billion monthly users. The average question asked inside it is three times the length of a traditional search, and more than one in six US users now include inputs beyond text, such as a photo or a voice recording. Google Lens, which searches from a photo, handles more than 25 billion searches a month, so a facilities manager can now photograph a worn part and ask who supplies it.

AI search engines pull answers from any format they can parse, so a video can get cited just as easily as a web page. Meltwater found YouTube was the most-cited platform across eight AI products in August, while Peec found Google’s AI Mode building answers directly from videos’ automatic transcripts. And the mix changes fast. Tinuiti and Profound looked at commercial searches and found YouTube’s share of AI Mode citations more than quadrupled between January and April. If you only publish in one format, you could lose visibility in a single quarter.

Why accessibility is the baseline for AI visibility

The accessibility layer that helps people with disabilities use your site is the same one AI agents rely on to read it. When buttons and menus are properly tagged in your site’s underlying code with what they are and what they do, screen readers can navigate them, and OpenAI confirms its agent in the Atlas browser relies on those same tags. Alt text tells Google what an image shows, and a video’s description and transcript give AI text it can quote directly. The quickest way to check all this is Google’s free Lighthouse tool. It flags missing alt text and unlabelled buttons, and since June it also scores how easily an AI agent can navigate your page.

The Pebble homepage, showing an image of a scenic road with a van driving on it pulling an electric RV.
Screenshot of the Accessibility Tree in Chrome Developer Tools, showing a simplified, parallel representation of a web page.

Why buyer personas and brand feel decide who gets chosen

Accessibility helps AI understand your site, but it doesn’t give a buyer a reason to pick you. For that, you need to know who’s searching. A head of operations and a finance director can ask about the same service in their own words and get answers built from completely different searches. Since last year, ChatGPT has also drawn on a person’s past chats to tailor its responses. Because every buyer’s answer is assembled from a different set of pieces, testing a single list of keywords no longer works. In our AI visibility work, we build test prompts from a client’s real search queries and live-chat transcripts, then group them by persona.

An AI summary might get your facts right, but it will flatten your tone in the process. Your images and video can carry your brand through, whether it’s a product shot in a carousel or a short clip of your team explaining how you work. That’s often the first impression a buyer gets of you, so treat visuals as carefully as your copy, and give them clear alt text and descriptions so search engines know what they show.

When each asset might be experienced in isolation, visual design should not just be about how a whole page looks together. Defining a content strategy first on our projects helps us understand what each audience needs to see and hear, with our design and build work amplifying the core site messages.

What to do now

  • Firstly, audit your templates by running Lighthouse’s accessibility and agentic browsing audits on your main page templates and fix what they flag. Start with alt text on images and accurate transcripts on videos.
  • For each thing you sell, list who asks about it and how they phrase the question. Then check your site answers it in plain HTML text on its own page, not buried in a PDF or saved for a sales call.
  • Next, ask those questions in AI Mode and ChatGPT the way each persona would, including by voice and with a photo. Note what the answers leave out and how they describe you.
  • Finally, check which images and clips AI answers pull in, and ask whether they’d still feel like your brand if a buyer saw nothing else.

AI now reads your site in pieces, so each piece needs to get your facts and your brand across by itself.

If you’d like help with any of this, our content strategy and UX work goes further than these steps. We set the messaging and content foundations for each of the people you need to reach, then design and build around them, so your brand holds up wherever your buyers meet it.