AI Mode (SEO is dead... again), New Info on AI Content from Google, Google Discover Desktop Rollout
SEO TL;DR #72 27/05/2025
Welcome to the 72nd edition of SEO TL;DR – a day late thanks to the British bank holiday, which I spent running around Flamingo Land (theme park #4 this year). A belated happy Memorial Day to my US readers too.
This week, I sum up the thoughts on AI Mode and its inevitable rollout, with news straight from Google on how to get included. I also share some hands-on impressions of Google Discover on desktop - the only thing which currently counters Google’s insistence on keeping all (non-paid) traffic to themselves.
AI X SEO
Google’s AI Mode Live in the US
As announced at Google I/O, AI Mode has now rolled out to all US searchers. It’s housed in its own tab (for now), but VP of Search Hema Budaraju confirmed this is “the future of Google Search” - and myself and others think AI Mode will quickly become the default.
Not in the US? Switch your VPN to the US and go to google.com/aimode
Powered by a custom version of Gemini 2.5, AI Mode introduces features like:
Deep Search for expert-level answers
Live Search with Lens integration
Personalisation using Gmail data
Custom Charts built on real-time data
Agentic Shopping with price tracking and Google-assisted checkout
👉 See Barry Schwartz’s full breakdown for all the new features and reactions.
It’s also been confirmed that AI Mode reporting will be coming to Google Search Console 🙌 I would have bet money this would never happen (where are you AIO data), but I’m happy to have lost that bet.
💡 Takeaway: A longer takeaway than usual and potentially warrants its own blog, but I can’t understate how much of a fundamental shift AI Mode is in how Google Search works, how content is discovered and the future of SEO.
With personalised SERPs, agentic features and AI Overviews becoming the default experience, traditional SEO levers like rankings and organic CTR are losing their predictive value. The same query now produces wildly different results based on personal context, meaning “position 1” no longer guarantees traffic (or even visibility).
For SEOs, this demands a massive mindset shift…
Content strategy must mature. Informational articles that once hoarded traffic through word count and backlinks are now being answered directly by AI search. Content needs to offer value that AI can’t synthesise with strong brand affinity, lived experiences and trust built across multiple channels.
Your brand must exist beyond the SERP. Citations in AI Overviews don’t carry the same weight as a click to your site. Visibility now depends on how well AI models understand and trust your brand, which means building a recognisable presence. Google have said many times now that brand is the future of search.
Link building isn’t dead. The importance of link building is something that Google has downplayed over the years, but we now know that links help establish your brand as a reliable entity in the AI model’s knowledge graph. Think less about DA and more about context, credibility and relevance.
The Messy Middle has become the invisible middle: With AI now able to complete tasks such as finding products, booking tickets and making purchases, entire user journeys can unfold without your website ever being visited by a human. This could mean less behavioural data for marketers and more reliance on structured data and product feeds to fuel these agents.
Reporting needs to evolve. Traditional metrics, such as how many keywords you rank for and the clicks they receive, are becoming more of a vanity metric. Traffic drops are inevitable, but they now need to be framed in terms of user intent, conversion quality and business outcomes. You’ll need to explain why rank tracking and impression data may no longer reflect real-world performance.
AI Mode signals the start of a more closed and fragmented search landscape, where the most visible brands will be those the model already trusts, not necessarily those with the best-optimised content.
If you're still optimising for rankings alone, you're playing an old game. The new SEO challenge is to exist within the AI-driven ecosystem, not just to be found, but to be understood.
Content SEO
Google Discover Rolling Out on Desktop
Google Discover is starting to appear on desktop for some users, including in New Zealand and other non-US regions, as reported in Search Engine Roundtable.
I can also see this in the wild by switching my VPN to New Zealand. The feed was on point in fairness, mirroring the very topics in this newsletter… complete with ✨Sponsored Ads✨
You can turn it off in preferences, which is the first thing I did, as it’s a major distraction for someone who Googles as much as I do.
💡 Takeaway: Discover on desktop could provide a new organic traffic source for publishers, much needed with AIOs and AI Mode. It’s worth revisiting your content for E-E-A-T and engagement signals, as Discover is notoriously opaque and driven more by user interests than keyword targeting.
Visibility here may not be SEO in the traditional sense, but optimising for discovery and shareability still applies.
AI X SEO
New Google Docs on AI Content & Search Features
Google quietly released two new help pages that clarify how AI Overviews and AI Mode work, and what site owners should know.
The pages break down how AI Overviews and AI Mode surface content; the TL;DR…
They use a “query fan-out” to issue sub-queries and find diverse supporting links
There’s no special markup or technical tweak required to be featured
Standard best practices still apply (crawlability, internal linking, structured data, up-to-date business profiles, etc.)
There are no additional technical requirements beyond what’s already expected for classic Search
AI-generated content isn't banned, but it must meet quality, accuracy and originality standards
Repeating the usual refrain: if content is low-effort or offers no added value, it's likely to be treated as spam
Google encourages transparency about how content was created (e.g. disclosing automation and using metadata)
For ecommerce, AI-generated product info must be correctly tagged using the “TrainedAlgorithmicMedia” attributes
💡 Takeaway: For now, Google’s AI systems still rely heavily on classic indexing and SEO signals. That means your existing SEO efforts haven’t been wasted - but now there’s a twist. You’re not just optimising for humans but also for how AI models interpret your content.
Now, this isn’t ever something I would recommend, but if you solely wanted to optimise for the robots, the aim is clarity, structure, and authority.
Structure for clarity using Q&A blocks, numbered lists and bold summaries so LLMs can extract and summarise your content cleanly.
Mention your full product, brand and person names. Avoid vague pronouns or flowery language AI might skip
Publish where LLMs look. Share content on Reddit, Quora, Medium and other authoritative sites, frequently cited in AI Overviews
Lean into Schema, such as author and organisation Schema and link between related articles to establish topical depth
Label AI-generated content properly and be transparent about how it was produced
There’s nothing new here except for the emphasis: Google is making it clear that your site’s future visibility depends not just on ranking, but on how well your content supports AI-generated answers - even if you’re just a citation.
Ecommerce SEO
Shopify Summer ’25 Edition Drops: Horizon, AI Store Builder & More
Last week, Shopify dropped its Summer ’25 Edition with 150+ updates, headlined by Horizon - their new AI-powered theme that introduces reusable theme blocks and drag-and-drop flexibility without custom code.
You can now generate entire storefront sections with text prompts. At the same time, Sidekick also got smarter (and multilingual), and a new knowledge base app helps your content surface better in AI shopping results.
Other highlights include discount codes in the cart (finally), POS upgrades, store credit refunds via third-party apps and Shopify Payments expanding to 16 more countries.
#SEOcharity
SEO for SEOs Sake Doesn't Work - Here's What Does
Two weeks ago, I enjoyed sitting down with Anton Shulke and Olesia Korobka and geeking out about my specialist subject.
I argue that blindly chasing SEO wins without thinking about users or business goals is a fast track to irrelevance.
We’re long past the 10-blue-links era. Between AI Overviews, UGC-heavy results, and personalised SERPs, SEO is less about “ranking” and more about relevance, trust, and user experience. Google’s message is clear: be helpful, not manipulative.
🍿 Watch the full talk below and donate to SEOcharity if you can - it’s a great cause
The gist of what I went through:
Forum content is rising – Reddit, Quora, niche sites. It’s time we think beyond blog posts and consider where experience-driven content already lives.
Don't optimise in a vacuum – Chasing DA, stuffing schema, tweaking dates to look “fresh”... it doesn’t work if the user gets nothing from it.
Click-through still matters – Google might normalise click bias, but user engagement still sends signals. Craft titles and meta descriptions that actually appeal to humans.
Use AI wisely – Great for ideas and speed. Terrible at nuance, tone, and originality. AI won’t replace you - but someone using it well might.
Real trust signals win – Reviews, testimonials, product comparisons, improved site search, even internal linking that reflects how people actually shop.
My golden rule? If you’re only doing it for SEO, don’t. Good rankings should be a side effect of good SEO.
PS: If you want to talk about your next project, use the link below to book a 20-minute discovery call: