When an AI summary appears on Google, we visit traditional links only half as often as when no summary is shown. ChatGPT Search, Perplexity, and Google’s AI Overviews have completely reshaped the SERP. Discovery has shifted from “type keywords, click around” to “state the problem, get a referenced solution,” often without leaving the results page. The era of the zero-click search has officially arrived, and in a zero‑click world, being the answer is the strategy.
I can’t help but think how relatable this is. AI Search makes things a lot easier to get answers to, but I find myself definitely clicking on results less (though I still try to do to make sure AI isn’t hallucinating).
My traffic from Google to WordPress sites is going downward at the time. But to forum is increasing. I don’t know who is to blame and thank.
But that blog gave ultra-positive image how AIs work. Solution of Google is now at same level than theirs translator… But it is defenetly race of amount and popularity, not content and its accuracy.
And now I’ve started to be annoyed of using my content for 3rd party business — but that is worth of another topic, and not necessary here.
Blogging died when content creators realized they couldn’t make money from useless content that would entice anyone to return. So they moved to images and video, where the algorithm forces visitors to watch content – a format made so simple that even a goldfish can watch it. Short-form content is, in any case, the ultimate Western disposable culture, and is not intended for use via search.
By blogs, you probably mean copy&paste[1] sites. The ones that are a dime a dozen and can be recognized by headlines like “Top-5 Forum Software for Better SEO (2025)”. They are made solely for ad impressions, not for content. Thanks to AI, more and more of them are being created. No one wants to read them because everyone knows that the “information” they provide is useless. Forums used to offer a more authentic form of information in a way. But that is changing, and AI pollution has reached Discourse too. Questions based on using AI content as primary content are on the rise here.
But text content is doing well and can continue to live strong. Streaming didn’t kill books either, even though it sometimes feels like literacy is going out of fashion. I mean that Discourse is not very different from WordPress. Both are based on text, and most people use them similarly.
The difference is that WordPress is based on a single publication. Discourse is most often based on a question/answer series. The end result is always the same: a kind of hit-and-run. A forum generates discussion because it’s a forum, just as poorly as WordPress generates discussion because it has comments.
My forum doesn’t get a single click from search results because it’s Discourse. Clicks and growth come for two reasons:
the forum has enough useful information
CDCK has enabled better SEO, and the content is indexable at all
But – the traffic to my forum, and its growth, is significantly based on my way of creating WordPress-style introductions. I create an article-like introduction, which generates just as little discussion as if I had published it on WordPress.
Topics get less traffic via Google. Apparently, the structure “question - guess - request for clarification - more information - answer - second answer - alarm” is difficult for search engines (should we switch to a method where a summary or digest is made from such a thread, a new thread is created from it, and the original discussion is moved under it?).
So, Discourse has succeeded in making my forum’s content visible (a small language area and a narrow niche make it much easier).
But it doesn’t convert into discussants for me, nor naturally returning visitors. Of course, the forum has fulfilled its purpose in terms of search results if a random visitor has found what they were looking for. So, the problem with my forum is not actually search engine visibility, but the lack of the famous critical mass that generates and maintains discussion, which would provide Google with indexable content, which would bring new traffic in, which would convert at some low percentage into new discussants.
But that’s no longer a question related to the topic itself.
Who knows if this whole answer is even related to the topic. I just took the opportunity to test an AI translator with a sufficiently long text… yes, I’m a bad person
we are missing a notation to prevent translation of, for example, fairly established phrases ↩︎
I find your reply very interesting @Jagster, it hits many important points and trends.
Forever, the vast number of visitors were anonymous users who just want info and to move on with their day. This is not a new problem. Having a core set of users who care enough and belong to a community is the key to surviving over time. A forum with no activity does not survive.
A big challenge AI poses is a drain on attention; we all have limited amounts of attention. If 99% of our attention goes to robots, only 1% is left for other stuff. This trend can be a pull on your most active users and is certainly something to watch for. In the long term, does anyone have the ability to compete with robots that love us and say yes all the time? The fundamental thing we at Discourse bring to the table is human connection. We believe this is deeply important and think it will win.
All companies have a high priority to get the right answers out there. This article is talking about how “the new SEO” is about getting quality, correct information out there.
In an ideal world, if a company feeds a “hit and run” anonymous user with the correct information right away, this can help in two aspects:
Community content quality; asking 100% duplicate questions is very taxing on a community.
Support is a very expensive arm of most companies out there; this reduces costs of private support because nobody needs to open a support ticket.
I think a great unlocker we will have is tools like ask.discourse.com (RAG based on live community content), which help ground answers right away with the most recent up-to-date information. Running engines like this is a great way of routing people to the right place in the community and answering all the easy questions right away.
They are also great at helping community managers understand “SEO pathologies” your community has.
Wrong answers the robot gives are an opportunity to correct the backing information. Post more canonical content (in search-prioritized categories), or remove search landmines.
Missing information on the forum that was never discussed but people keep looking for is an opportunity to kick off discussion in topics lots of people care about.
The age of AI is presenting us with a ton of new and wonderful challenges:
AI just sucks all the great information out of communities and can feel parasitic to many communities. There is no scheme of licensing we can opt for. I have noticed a reasonable amount of private communities pop up and many places with gated content. An interesting approach is figuring out what content you should gate to enable a more thriving community?
As community engine providers, do we need to build new systems that “extract the summarized truth” out of a giant messy community and kill off the lies? Then give the “golden information” to search engines and LLMs?
People are increasingly using AI to author a community with robot-authored content that is being presented as human, which is dangerous. Trust is paramount; this is why we believe strongly in ALWAYS labeling AI-generated content.
How do we remind people of the value of human connection? A “join the conversation and talk to humans” CTA starts being an interesting thing to explore in the age of robots. Discourse ID can lower the barrier of joining.
At the core of everything though, great content with a thriving community are the biggest ingredients you need for success. Start there, then lean on the many Discourse features to make it more effective (docs/wikis/search priority/reporting/Q&A/solved/ideation, and more).
