I’m seeing a really frustrating stats with our community forum activity, which is declining every month. Our forum is a sort of a complement to our SaaS service, where users can get support, leave feature requests, follow updates in the changelog, check out CSS snippets and pro tips, etc.
Is it just our specific forum, or are you noticing something similar? Are you going to shift gears and test out other approaches, or are you maybe thinking about shutting down the whole project?
Sure! Here are some numbers on users’ topics and comments. We had a sudden peak a year ago so probably we should not inlude it in the stats at all, but the results were great:
The confusing part is that some metrcis look ok, but the involvement is clearly decreasing. I’m trying to figure out the root of the problem and my next steps.
Hmm… is the content very predictable, e.g. are the same kind of queries always popping up? Are there active, debatable discussions that users can participate in?
Speculation… but I’ve been wondering if more people becoming more comfortable turning to AI as a first port of call would impact communities (especially those with a support element).
I’m tentatively thinking that not only would there be fewer questions being asked of the community, but also fewer opportunities for existing community members to answer them. Less to read also may rattle down to less reason to linger. Plus if there are fewer ‘easy questions’ then it’s harder for newer members to convert from ‘askers’ to ‘answerers’.
I had the exact same thought but didn’t want to point to AI as the easy culprit right away but… it does make sense
On the bright side, that may mean it will rectify itself over time when the AIs have no “new content” to scrape, returning no useful info, and people flow back to the good ole’ forums.
We’ve also seen a steady downward trend in engagement - specifically around the 1:many peer support areas of our community.
In some ways, it’s good to know we have strong support documentation, and equally strong and accessible community discussions that enable AI to surface information to users in a way that solves their questions at the ever increasing speed they’re growing to expect.
I do think though that AI is changing the way that support based communities operate. And the ways those types of community show up is beginning to shift.
I know we’re consistently asking ourselves now “what are the things we do that AI can’t?”, like the engagement that goes beyond support (e.g. events, activities). Why would people return to the community now that one of our main habitual drivers is now being replaced? What becomes the new driver for people to habitually return to our community spaces?
There’s a potential information degredation cycle here to look out for too: if communities go away because people are going to AI first, AI also loses community as a source of information.
At that point AI will become worse at answering new questions about your subject matter, and people won’t have another place to go to ask either.
Right?! I think there’s a lot of creative angles that can be taken to supplement the traditional “ask a question, wait for the community to respond” approach, which in turn provides new reasons for community members to continue returning, while also providing new and relevant info for AI.
Just one example format that comes to mind (not fully developed, there’s probably more thinking to put behind it) is a “community feedback cycle”. Asking your community “What are things you struggle with, and often have to find creative workarounds for? What are those workarounds?”
It’s multi-angle, giving product teams insights into pain points, giving the community space to feel they’re valued and listened to, and gives AI insights into solutions people are currently using.
We’re certainly living in interesting times and being challenged to challenge the status quo!
Yeah, this is exactly what I’ve been thinking about as well - AI has already changed user behaviour, and whenever we need help our go-to place is AI (let’s face it).
But while it works great with general things, it probably won’t work when it comes to questions about specific service where only a dedicated support team coud help and provide a decent solution. That’s why I thought a support community still could have a chance.
ChatGPT advises testing out a light shift from a support forum to a place where users could find practical solutions from real people. Worth trying per se, but only if it makes any sense given the current situation with AI and stuff
Just wanted to find out whether someone else is experiencing the same issue.
My forum is more than 30 years old and about a quite niche activity.
It has steady activity and engagement, but I think the members grow older and “retiring” users are not completely replaced by new folks. I might see a decrease in several years, but who knows I’m not sure AI has a real impact on my forum health, but I guess it depends from a forum to another.
I sure hope that in the future forums will spark again the same interest in people as they did in 2005-2010
I think it’s always interesting to share this curve: from 2001 to 2026 (pre-2001 data was non-existent), 15 years of activity. Sharp spike in mid-2020 was migration to Discourse… .
I would not have tought about that, but that happens to me too. For many questions, AI has unfortunately been way faster and reliable than communities to answer .
I wonder what could be done for a hobbyists[1] forum to be more attractive and more appealing. More than AI; additionally to more than social media.
For me, I think there are also Discourse forums with large communities, and doing updates is too risky (unmaintained plugins, same for components, etc.), so as long as their forums work, they don’t take risks. Plus, there are plenty of software options nowadays; users no longer want to deal with complicated setups like Discord, Facebook, WhatsApp, Matrix, etc. They just pick a channel, ask their question, and boom—there’s the answer. They don’t even want to search, even if their question has already been asked multiple times in the discussion channel. I think mindsets are changing, and it’s just so much easier to set up. There are also all the providers or manufacturers who publish links on their chat apps: “If you have a problem, join us on Discord, Facebook,” and so on. The community has changed a lot. And yet, Discourse has evolved a lot; for me, Discourse is a bit like a sandbox—we have a base and adapt it according to our community, trying, in my case, to simplify usage while keeping a structured forum. Maybe the design of Discourse is a bit outdated compared to what we see now, but we’ll see with future themes like Meta, but it might be too late. I don’t understand people; every day, I have to log into the Discourse community forum—it’s my little morning drug to see additions, patterns, and especially ideas . After all, is the number of forums increasing or decreasing? And then, I agree with you regarding AI. To sum up, if we could see new faces on the forum, I’d be happy .
but I’ve been wondering if more people becoming more comfortable turning to AI as a first port of call would impact communities (especially those with a support element).
.
.
. After all, is the number of forums increasing or decreasing? And then, I agree with you regarding AI. To sum up, if we could see new faces on the forum, I’d be happy 
.