AI Search News | Australia | 1 November 2026
Last updated:
Google updated two Search Central documentation pages in the last ten days: its guide to preventing user-generated spam (26 October 2026) and its URL structure best practices (23 October 2026). For Australian online stores, forums and booking sites, both changes point to the same job: control which URLs Google can index.
This is a documentation story, not a product launch, so we are labelling it honestly. No new Google feature or ranking update was announced by Google in the past seven days that we have not already covered, so this edition uses the third option in our editorial order: the newest official Search Central change, framed around what we saw in Google AI Mode in three Australian cities this week.
Key Takeaways
- Google updated its user-generated spam guide on 26 October 2026 and its URL structure best practices on 23 October 2026.
- The spam guide tells platforms to apply noindex or block crawling in robots.txt for pages meant for sharing, not public search discovery.
- Google recommends hyphens over underscores in URLs, and warns that non-compliant URLs can cause extremely high crawl rates or none at all.
- A spot check of Google AI Mode across Brisbane, Perth and Adelaide on 1 November 2026 found it relies on star ratings, review counts and years in business.
- Neither update is a ranking or product change; Google confirms both are documentation refreshes with no stated country restriction.
What happened
Google keeps a dated log of every Search Central documentation change at developers.google.com/search/updates. Two entries from the last ten days matter to Australian site owners.
26 October 2026: updated guidance on preventing user-generated spam. Google’s log says it updated the user-generated spam prevention guide “with guidance on blocking indexing of link-shared content and monitoring abuse in Search Console.” Google’s stated reason: “To help platforms keep user-created pages that aren’t meant for Search out of the index and detect spam earlier.”
We opened the guide itself. It now tells site owners and platforms to think about whether content is meant to be found through search or only reached by someone holding a link. In its words, when pages are created “for direct sharing, collaboration, or utility rather than public search discovery,” the guide says to apply the noindex rule, using a robots meta tag or an X-Robots-Tag HTTP response header, or block crawling in robots.txt by default.
The guide also has a monitoring section. It recommends verifying your site in Search Console so you get email alerts about manual actions and security issues, and checking the Page Indexing report and the Search performance report regularly for “unexpected spikes in indexed pages, impressions, or queries that don’t match your platform’s purpose.” It suggests tracking platform-level creation and indexing metrics and setting alerts for anomalies, such as spikes in newly published public pages.
The older advice is still there too: publish a clear abuse policy, consider noindex on posts from brand new users with no reputation, add nofollow or ugc rel attributes to links in untrusted content, use moderation for suspicious interactions, block automated account creation with reCAPTCHA or similar tools, and keep an eye on server logs for sudden traffic spikes.
23 October 2026: updated URL structure best practices. The log says Google updated the page “to clarify word separation guidelines using hyphens versus underscores,” and gives the reason as synchronising the documentation with the episodes of its “The Non-Canonical Cabinet” video series. The page recommends hyphens to separate words in URLs, and explains that “for historical reasons, we don’t recommend using underscores.” It also says joining words together, such as greendress, is not recommended.
The same page lists requirements for a crawlable URL structure. URLs should follow the IETF STD 66 standard, URL fragments should not be used to change page content (use the History API instead), and URL parameters should use the common encoding of an equals sign between key and value and an ampersand between parameters. Google warns that if URLs do not meet the criteria, Search “will likely crawl your site inefficiently,” including “extremely high crawl rates, or not at all.”
Neither change introduces a new rule. Google describes the 21 October update to its control what you share page the same way: a refresh to match how the company already explains things. We covered that in Google Search Controls Update: What Australian Sites Need. Taken together, the three October changes show a consistent theme: Google wants site owners to decide deliberately what is indexable.

Is it live in Australia?
Documentation is not rolled out by country. These pages sit on Google’s global developer site and the two entries in the updates log do not mention any country or region restriction. Google has not published anything saying the guidance applies differently in Australia, so Australian sites should treat it as applying to them, the same way it applies everywhere else.
What Google has not said is whether any of this changes how Search or AI Mode behaves. The log describes both entries as documentation updates, and the stated reasons are to keep pages fresh and in sync with Google’s presentations and videos. We are not claiming a ranking change, a crawling change or an AI Mode change, because the sources do not claim one.
What we saw in Australian AI search this week
Because the story is about which pages get found, we ran our own national spot check on an industry that lives and dies by URL structure: online stores. We asked Google AI Mode one buying prompt in three cities, “best ecommerce store in” each city. The instrument is Google AI Mode, one query per city, run on 1 November 2026 (AEST). It is a single query per city, so treat it as a snapshot, not a measurement. We did not test ChatGPT or Perplexity.
| City | Businesses named | Signals quoted |
|---|---|---|
| Brisbane | Our parser counted 1; the visible text names at least 6 (four web agencies and a mix of local online retailers) | No star ratings quoted; one cited source (a startup directory) |
| Perth | 8 counted, almost all web design and Shopify or WooCommerce agencies | Google star ratings and review counts shown for the agencies, plus awards and years in business |
| Adelaide | 7 counted, a mix of e-commerce agencies and local online retailers | Google ratings and review counts, years in business, one agency website cited alongside Google |
Two observations stand out. First, the prompt “best ecommerce store” was read by AI Mode as two different questions. In Brisbane the answer opened by saying the result “depends entirely” on whether you want the developers who build online stores or the local brands that run them, then split into two lists. In Perth and Adelaide the answers leaned heavily towards agencies. For a store owner, that is a reminder that AI Mode interprets intent before it ranks anyone, and vague category language can send you into the wrong list.
Second, the signals AI Mode surfaced in Perth and Adelaide were the kind that come from a business’s Google profile and website: star ratings, review counts, a services description and years in business. Brisbane showed none of the rating signals and cited a single directory. Our parser also undercounts when the answer is long, because we only capture an excerpt, so the Brisbane figure in the table is a floor, not a count. We report these numbers as they came back and make no claim about any individual business.
None of this is caused by Google’s documentation change, and we are not suggesting it is. We include it because it shows what AI Mode draws on when it recommends an online business: clear service descriptions, consistent profile data and pages it can reach. Those pages have to be indexable and have sensible URLs first.

What it means for Australian businesses
This is our view, not Google’s. The documentation changes are small, but they land on problems we see on Australian websites every week, in every capital city and in regional areas.
Online stores in Sydney, Melbourne and Brisbane. Filter and sort pages, search result pages and cart or wishlist URLs can generate thousands of near-identical URLs. Google’s URL guidance on parameters and its warning about inefficient crawling are directly relevant. A store with a tidy structure and a clear policy on which parameter pages are indexable gives crawlers, and therefore AI answer engines that rely on indexed pages, a cleaner picture of the catalogue. Our Shopify website design guide covers how we set this up on a platform that generates many URLs by default.
Community and membership sites in Perth and Adelaide. Any site that lets visitors post, such as a local forum, a classifieds board, a review section or a Facebook-style group page, is the audience for the user-generated spam guide. The new emphasis on monitoring is practical: a sudden spike in indexed pages or in impressions for queries unrelated to your purpose is often the first sign someone is abusing an open form.
Tool and booking platforms. The guide specifically mentions services where users “publish shared pages, generated content, or web apps.” Many Australian software start-ups let customers create quote pages, shareable proposals, booking confirmations or client portals. Those pages are useful to the person holding the link and useless to Google Search, so they should default to noindex. If they are indexed, they can bloat your site, expose information you did not intend to publish and attract spammers.
Regional service businesses. A tradie in regional Queensland or a vineyard in the Barossa may not run a platform at all, but most small business sites on WordPress still have an open comment form, a contact form that auto-publishes anywhere, or tag and archive pages that were never meant to rank. The principle is the same: decide what should be found, and keep everything else out of the index.
The AI search connection. AI answers draw on indexed, trustworthy web content. A site whose index footprint is full of thin tag pages, spam comments or accidental duplicate URLs gives answer engines more noise to sift through. A clean, deliberate footprint is a precondition for being quoted. That is why we treat index control as part of answer engine optimisation and generative engine optimisation, not only as technical SEO.
Is your index footprint helping or hurting you?
Google’s latest guidance asks you to decide which URLs belong in Search. Our Free AI Readiness Check looks at what AI search and Google can actually reach on your site, and shows you what to fix first. It takes about 15 minutes.

What to do now
- List every page type that visitors can create or that your platform generates. Comments, profiles, shared links, generated outputs, filter pages, internal search results and confirmation pages all count. Decide for each one whether it is meant for public search.
- Apply noindex to pages meant only for people with the link. Use the robots meta tag or the X-Robots-Tag header, as Google’s guide describes. Remember that blocking in robots.txt stops crawling but a noindex tag can only work if Google can read the page, so choose the method deliberately for each page type.
- Add rel=”ugc” or rel=”nofollow” to links in untrusted content. This is the long-standing advice in the same guide and it removes the main reason spammers target open forms.
- Verify the site in Search Console and check the Page Indexing report monthly. Look for sudden jumps in indexed pages. Google’s help centre explains the report in its Page Indexing report guide. Set alerts so you hear about manual actions by email.
- Audit your URLs for underscores, fragments and unusual parameter encoding. Switch to hyphens for new URLs. Do not rewrite old URLs without proper 301 redirects, because changing a working URL has its own risks.
- Keep your business profile data consistent. Our spot check shows AI Mode drawing on ratings, review counts and service descriptions. Make sure your name, services and structured data, for example LocalBusiness markup, say the same thing everywhere.
If you would like this checked for you, our SEO team reviews indexing, URL structure and crawl waste as part of every audit, and our website care plans keep spam and indexing issues under watch month to month.

How this fits the bigger picture
Over the past month we have tracked a steady run of official changes: the UGC Fresh Data Program for platforms with fast-changing user content, the September 2026 spam update, and now the guidance on keeping non-search pages out of the index. Read together, they describe a search engine that rewards sites which are deliberate about what they publish, and that wants platforms to police their own user content.
For Australian businesses, the practical conclusion is calm and unglamorous. You do not need to rebuild anything because of a documentation update. You do need to know what your site is publishing to the index, because AI search can only quote what it can find, and spam or accidental duplicates make that harder. For a wider view of the market, see our AI Search News hub, and for the longer view on how AI search is changing local discovery, our analysis of whether AI search is just Google.
Frequently Asked Questions
Did Google announce a new spam update on 26 October 2026?
No. The 26 October entry in Google’s Search Central updates log is a documentation change to the user-generated spam prevention guide. Google’s Search Status Dashboard most recently lists the September 2026 spam update as complete. Always check the Search Status Dashboard for ranking updates.
Should I use noindex or robots.txt to keep shared pages out of Google?
Google’s guide mentions both: apply the noindex rule using a robots meta tag or X-Robots-Tag header, or block crawling in robots.txt by default. They work differently. Noindex lets Google crawl the page and then keep it out of the index, while a robots.txt block stops the crawl. Which is right depends on the page type, so decide per section of your site.
Do I need to change underscores to hyphens in my existing URLs?
Google recommends hyphens for new URLs, but the documentation does not tell you to rewrite working URLs. If you do change an existing URL, set up a permanent redirect from the old address to the new one and update your internal links.
Does this affect whether my business appears in Google AI Mode?
Google has not said it does. The updates are documentation changes. Our own spot check suggests AI Mode draws on business profile signals and website content, so keeping your index footprint clean and your profile data consistent is sensible, but we cannot show a direct link from these two changes to AI Mode results.
Which Australian sites should care most?
Any site with open forms, user accounts, shareable generated pages or a large online store catalogue. Smaller brochure sites should still check comments, tag archives and any auto-generated pages.
Make Titan Blue a preferred source in Google
One click tells Google to surface our articles more often in Top Stories, AI Overviews and AI Mode.