Google launches AI mode chatbot in search

[Whitey]

Just announced:
[theguardian.com...]

www.theguardian.com Google unveils -˜AI Mode' in the next phase of its journey to change search Search engine revamp and Gemini 2.5 introduced at conference in latest showing tech giant is all in on AI

[Whitey]

Some more food for thought: Analysts warn Google’s search share may drop below 50% in five years, threatening the traffic lifeline many smaller websites rely on. With AI answers reducing clicks and users shifting to social and app-based discovery, now’s the time for smaller publishers to diversify traffic and build direct user relationships.
[reuters.com...]

[Whitey]

And more:
[seroundtable.com...]

[lucy24]

launch’s?

[bgweb]

This looks like a huge change for publishers. It's very difficult to imagine that a lot won't go to the wall.

[Whitey]

@lucy24 – Good catch! One typo, now immortalized.

@bgweb – You’re right, this shift could be existential for many. But there are levers that can still be pulled. Based on recent audits I’ve observed, here’s a quick survival guide for smaller sites:

1.Diversify traffic sources – Don’t wait. Build email capture, push notifications, social visibility, and partnerships. Google is no longer the only gateway to discovery.

2.Lean into brand – Even modest brand recognition helps. Navigational queries (e.g. your brand + topic) can lock in visibility via NavBoost.

3.Create content AI can’t easily replicate – Focus on original data, opinion, comparison, and location-specific insight—areas where generic AI summaries fall short.

4.Improve behavioral signals – Fast load times, strong mobile UX, scroll depth, return visits. These matter more than ever when search is driven by engagement signals.

5.Prepare for zero-click answers – Structure your content for visibility in AI summaries. That means schema, scannable answers, branded visuals, and clear authorship to increase the chance of attribution—where the AI credits or links to your content in its response.

6.Test and build retention – Small tests on loyalty features, saved searches, or personalized follow-ups can turn passive visitors into repeat users.

The rules are changing fast—but that also means new opportunities for those agile enough to adapt.

[bgweb]

@Whitey - that's very useful.

One thing I’m wondering is whether we’re starting to see a kind of “digital inbreeding.” Large language models must now be feeding off content they’ve generated themselves, warts and all. As a result, the potential for incorrect information to be amplified increases. This trend is likely to continue as more small publishers either switch to partial or full subscription models, or shut down entirely.

In an ideal world the owners of LLMs would issue tokens to publishers based on how much of their content is used. These tokens could then be cashed in or used to purchase access to the AI APIs themselves.

[Whitey]

@bgweb – Great point, and yes, self-reinforcing loops in LLMs could degrade quality over time. The token idea is smart. Maybe what we need is a Content Attribution Protocol, like a “robots.txt for LLMs”, that registers, tracks, and pays. Might kick off a whole new layer of web infrastructure. Anyone seen early versions of this being tested?

[phranque]

launch’s?

@lucy24 – Good catch! One typo, now immortalized.

now edited...

[Rlilly]

SEO Is Dead, According to GoogleGoogle is all in on Google AI.
[inc.com...]

[bgweb]

I'm getting a lot of ad agencies contacting me at the moment. I'm guessing they're fully aware the pie is becoming smaller - a lot are likely to go to the wall.