SEO work is getting split across Search, images, and AI Mode

TL;DR: Treat Google visibility as a portfolio now, not one ranking report, because spam updates, image search data, and AI Mode reporting all require different fixes.

What actually changed for SEO teams?

Search Engine Journal’s “Google Spam Update, Image Search Data In GSC – SEO Pulse” by Matt G. Southern reports three things worth separating: Google’s September spam update may take up to two weeks, Search Console is adding image-based search data, and reporting is looking at traffic from Search and AI Mode.

That is not one story. It is three operating signals.

The spam update is about quality enforcement. If traffic moves during the rollout window, the first job is not panic editing. It is diagnosis. Which pages moved? Were they thin, copied, over-optimized, expired, doorway-like, or built mostly for search engines instead of users? Search Engine Journal reported the update may take up to two weeks, so reading day-one volatility like a final verdict is usually sloppy.

The Search Console image-data item is different. That points to measurement. Images have been a real discovery surface for years, but most SEO audits still treat them as decoration. Product photos, diagrams, screenshots, charts, logos, author headshots, and featured images can all influence how a page is found and understood. If Google gives site owners more image-based search data in Search Console, SEOs get a better way to audit visual assets instead of guessing from page-level traffic.

The AI Mode reporting item is the most interesting, and the easiest to overstate. Search Engine Journal says reports look at traffic from Search and AI Mode. That does not mean AI Mode is suddenly the whole game. It means the reporting layer is starting to reflect a search experience that is no longer just ten blue links, snippets, and ads.

What should you measure differently?

The old SEO dashboard usually starts with rankings, clicks, impressions, CTR, and conversions. Still useful. Not enough.

For a site that depends on search, I would split the audit into three lanes: page visibility, visual visibility, and answer visibility. Page visibility is the classic work: query intent, internal links, crawlability, content quality, technical health. Visual visibility asks whether images are original, useful, compressed, descriptive, placed near relevant copy, and worth indexing. Answer visibility asks whether a page has clear entities, factual structure, first-hand detail, and concise passages that can be cited or summarized by AI search products.

That last part matters because AI search often rewards extractable clarity. Not keyword stuffing. Not FAQ spam. Clear claims. Named people. Dates. steps. tradeoffs. Original evidence. If your page says the same generic thing as 400 competitors, an AI answer has little reason to care where it came from.

Ashe runs Lucky Domains, which includes SEO and search visibility work, and this is exactly where I would separate brand, content, and technical diagnostics instead of lumping all traffic loss into “the algorithm.”

What is the catch with AI Mode traffic?

The catch is attribution.

If AI Mode changes how people search, the visible click may become only one part of the value. A user might see your brand in an AI-generated answer and search you later. Or they might get the answer without clicking. Or your content might influence a summary without any visit you can tie back cleanly. That makes classic last-click SEO reporting weaker.

But weak measurement is not permission to hand-wave. Builders still need pages that deserve to be found, images that carry information, and analytics that separate surfaces. When a spam update rolls out, wait for enough data before declaring a winner or loser. When image data appears, audit the assets that already rank or nearly rank. When AI Mode reporting becomes visible, compare those visits against normal Search instead of blending them together.

Practically, I would start with the top 20 organic landing pages, then map each one to three questions: does the page satisfy the human query better than competitors, do its images add searchable value, and can its key facts be quoted accurately by an AI system? The catch most readers miss: AI search does not remove SEO discipline. It punishes lazy SEO faster, because vague pages, stock images, and recycled answers give machines nothing distinct to retrieve.