We’ve spent months talking about community as a growth channel. September ended with a slightly less comfortable question: can platforms actually prove their safeguards and automated decisions work? That’s where this week’s community management news gets interesting.
Going into October, Brazil’s ANPD enforcement team kept Discord's live-video suspension in place after deciding the evidence supplied wasn't enough to demonstrate that the required safeguards worked across the platform. The remaining appeal issues are still under review ahead of a final ANPD board decision.
Elsewhere, Reddit is tightening its grip on disguised marketing and changing how moderators and brands connect with the platform. Khoros has added better moderation reporting, external knowledge search, and automation histories. Those stories look separate until you put them beside each other.
Community operators have more moderation tools, but they're also being asked for better evidence of what those tools did. And October won't give teams much breathing room. Black Friday planning is underway, while many moderation teams are already stretched.
An October 7 Gainsight practitioner discussion captured the human side. One operator described keeping a former Khoros community running during migration, while another contributor, Linsey M, warned:
“It’s worth weighing the benefit of uninterrupted activity against the additional work it creates for the migration and community teams.”
It's a useful reminder that keeping communities running takes more than new platform features. Someone still has to manage the work when things get complicated.
TL;DR
-
Discord video is still suspended in Brazil under the latest available ANPD position. The regulator says restoration remains tied to evidence that the required safeguards work, while the remaining appeal issues are still being analyzed before a final board decision.
-
Reddit is putting more structure around how community operations connect to other tools. New public API access requests close October 31, RSS support ends November 13, and moderators can route some alerts into Discord or Slack through Discord Relay. An October 6 update also added tighter CRM workflows for brands.
-
Community moderation tools are getting more inspectable and more connected. Khoros Communities 26.5 improved reporting around AI moderation, added fuller Orchestrator execution histories, and introduced early-access federated search across sources such as Confluence. For buyers, the question is shifting toward whether a team can reconstruct what an automated system did and pull in the knowledge needed to fix the issue.
-
October is already the first live test for holiday community support. Adobe expects $95.8 billion in US online spending this month, up 8% year on year. Higher Logic's October 14 benchmark briefing may add fresh context, while Gainsight Pulse Europe follows October 19–20.
Is Discord Video Still Banned in Brazil?
As of the latest ANPD position available right now, yes. Discord's Go Live and equivalent video-sharing features remain suspended in Brazil, and the argument has moved well beyond whether Discord says its safeguards work. The restriction started August 17, making October 17 its two-month mark if nothing changes.
Discord's broader availability isn't affected, though: text chat and other non-video features can continue. The ANPD says restoration depends on Discord demonstrating that its safeguards work, with the remaining appeal issues still under administrative review.
Discord reportedly handed over screenshots, user communications, and internal checks. The ANPD wanted something harder to argue with: technical evidence showing the restrictions actually worked across Discord’s infrastructure. The appeal is still being analyzed, with a final administrative decision due from the agency’s board once that process is complete.
One thing worth mentioning here. The ANPD restriction is an administrative measure; the separate federal civil action seeks R$500 million in collective damages, which is a claim, not an imposed fine. Brazil's government outlined the claim in August.
That makes the timing of Discord’s wider age-assurance rollout interesting. Discord says more than 90% of users won’t need to verify their age manually, with account signals doing most of the work.
Discord co-founder Stanislav Vishnevskiy summed up the company’s position neatly:
“Trust is earned through actions, not posts.”
Brazil’s response, effectively, is: prove it.
Why Should Community Buyers Care?
The pressure increased again on September 30, when a federal court ordered Discord to retain moderation records for at least six months, strengthen protections for younger users, and stop removed criminal communities or accounts from simply reappearing.
The court-ordered reporting and the ANPD appeal shouldn't be treated as the same proceeding. For operators, the practical issue is similar: can a platform retrieve moderation records quickly enough to explain a decision? Discord's Brazil support page still lists live video as unavailable, with no restoration date.
Australia’s eSafety Commissioner and Ofcom have since added to the same broader argument around platform accountability. In fact, on October 6, Ofcom opened an investigation into whether Meta properly assessed Instagram Instants before launch. No breach has been established, but the demand for evidence before and after a safety change is spreading.
There's another UK angle worth watching. Meta, TikTok and X have challenged Ofcom's requests for detailed moderation evidence, arguing that the demands go too far. Ofcom says it needs that information to assess whether online-safety rules are effective. It's a live dispute, however, not proof that any of the companies breached the law.
For buyers, this is where the story gets useful. A safety feature on a product page is one thing. Being able to show what happened when it was tested, challenged, or failed is another.
How is Reddit Changing Community Infrastructure?
Authentic recommendations are becoming more valuable, and Reddit is tightening the systems that access and moderate those conversations.
Reddit's August infrastructure plan now has firm deadlines. RSS support ends November 13, and moderators using RSS alerts are being directed toward Discord Relay, which sends alerts into private Discord or Slack channels. Teams should test those connections before the cutoff.
The October 8 changelog introduces another dilemma. Reddit is testing automatic notices for manually removed posts without stated reasons, alongside pauses on repetitive cross-community posting. Some moderators worry the notices will generate more appeals and modmail.
Reddit’s also testing AI-generated summaries behind trending labels to help users find related conversations. It's a limited October experiment rather than a full rollout, and it raises a useful moderation question: when discovery changes, who handles the extra attention the system sends to a community?
The API changes could be more disruptive. Reddit says unregistered apps may lose access from January 12, 2027, with remaining public API access ending in March. Discord Relay can replace some moderator alerts, but there's no replacement for RSS feeds covering communities a team doesn't moderate. If support monitoring depends on those feeds, audit it now.
Reddit's Home feed test cut visible suspected spam by 28%, while user reports fell 11%. It also identified 485 coordinated accounts behind more than 44,000 pieces of undisclosed promotional content.
Reddit expanded lead capture into CRM workflows on October 6, with HubSpot integration in open beta. That's a commercial connection rather than a replacement for support routing. We've covered the wider trust and shopping argument in our Reddit AI shopping analysis. Here, the concern is what happens when moderators' existing connections stop working.
Are AI Community Moderation Tools Becoming Easier to Audit?
Khoros’ latest release gets into the weeds in a way I quite like. It isn’t promising AI moderation will magically take work off everyone’s plate. It’s giving admins a better chance of figuring out what happened when automation goes wrong.
Communities 26.5, released September 23 to 24, added more detail around AI moderation decisions and bans. Khoros also fixed an attribution issue that could make an automated action appear under the post author’s name, while Orchestrator now keeps fuller execution histories.
Khoros says admins can:
“see exactly how their automations are performing with full execution history.”
That matters when a member questions a ban, a post disappears unexpectedly, or support needs to investigate before taking over.
The release also pushes community further into the service stack. Early-access Federated Search can pull knowledge from Confluence or REST APIs, while Khoros shipped Salesforce connector fixes.
Its health dashboard tracks connected sources' availability, response times, and consecutive failures, with permissions checked per viewer. That helps admins spot broken knowledge connections before customers hit dead ends. Khoros confirms the feature remains opt-in Aurora early access.
Discourse has a different version of the connected-support idea. Its September 2026.9 release lets operators make an AI conversation the forum homepage and offers an Ask AI admin dashboard and MCP server. The forum can help answer a question even when a customer never opens a conventional discussion thread. That's an option to evaluate, rather than evidence that every support question is resolved.
How Should Communities Prepare for Black Friday?
Reddit’s Mod Reserves program is a reminder that holiday planning often comes down to having enough people when things get messy. It gives communities experienced volunteer moderators during traffic spikes or staffing gaps. Reddit calls it “more hands on deck.”
Most enterprise community teams don’t have that luxury.
The CMX 2026 report says 34.3% of teams have one full-time employee, while another 37.2% have only two to five. Meanwhile, its public findings say 93% of community professionals use AI tools. These are survey findings, not a claim that AI has solved staffing. Adobe expects US online spending to reach $95.8 billion in October, up 8% year on year. Holiday traffic is already building.
Higher Logic's October 7 analysis also found 103% more average monthly logins among associations using job-board integrations. That's a correlation, rather than proof of causation, but it suggests connected member services deserve attention.
Five Things I’d Test Before the Holiday Surge
-
Overflow ownership: Who steps in when the queue gets ugly?
-
Escalation rules: When does a post move to support, legal, or security?
-
Automation under pressure: Can filters handle higher volumes without blocking legitimate requests?
-
Deflection content: Are the common delivery, account, and product questions answered, and can you tell when the answer actually helped?
-
Human override: Who can quickly reverse a bad moderation decision?
I'd also test one complete support handoff. If a community answer fails, can an agent see the original conversation, the case owner, and what's already been tried, or must the customer start again? Can the community team see when support resolves it?
That's the kind of problem I'd rather discover before Black Friday.
Why Is Community Demand Rising Faster Than Teams Can Prove Its Value?
Gainsight has an interesting number from Gartner's 2026 Market Guide: inquiries about customer communities rose 33% between April 2025 and March 2026, compared with the previous year. Gartner says B2B buyers are increasingly looking to their peers for advice, too.
But I'm less interested in how many people clicked on an answer than what happened afterward. If they ended up calling support anyway, I'm not convinced we've saved anyone much work. That's something I'd be checking before the holiday rush.
The CMX 2026 report exposes the awkward bit: community professionals increasingly credit their work with positive results, but far fewer can put a confident business number on it. I'd be wary of claiming a support saving when all we've measured is a click.
For community leaders, October’s pressure is pretty clear. If community is carrying more of the buying conversation, its reporting needs to catch up.
What Else Changed in Community Platforms This Week?
Gainsight made Skilljar AI Tutor generally available on October 8, using learners' course progress and supported Wistia videos to inform answers. It's another step toward connecting customer education with self-service, although complex issues still need human support. Gainsight's October 7 update also added gamification points to community user exports, making participation data easier to work with.
Bettermode set October 15 as the deadline for removing GraphQL alias support, citing reliability and security improvements. Only integrations using aliases are affected, so teams should probably check custom queries and middleware rather than assume broader disruption.
Online-safety scrutiny is spreading beyond Brazil. Ofcom is investigating whether Meta properly assessed Instagram Instants before launch. Meta says it completed the necessary checks and will cooperate. Separately, Meta, TikTok and X are challenging Ofcom's demands for safety data. Neither proceeding establishes a breach.
If your Discord community relies on verification bots, the October 4 Double Counter incident deserves a look. Attackers compromised one of the provider's servers and abused its bot token, according to the company. Have I Been Pwned lists approximately 275,000 exposed unique email addresses. Discord wasn't breached, but what could someone do with the permissions you've given your third-party bots?
Higher Logic also scheduled an October 14 benchmark briefing drawing on its 2026 research covering nearly 1,200 organizations. It's a useful reminder to examine whether connected community services actually improve engagement and support outcomes.
In Other News
A couple of other things to watch this week:
-
WhatsApp's Black Friday headache. Retailers have plenty of reasons to message customers, but how many of those messages did anyone actually ask for? In an October 9 interview, Infobip's Kim Johal discussed the risks of overdoing promotional messaging. Nobody's going to complain about a useful delivery update. Five reminders about the same discount might get a different reception. Community teams should be thinking about consent before November's campaigns get underway.
-
Are those visitors even people? Acoustic CEO Jon Ziglar raised an interesting problem in his October 8 interview with us. AI shopping assistants can browse products on someone's behalf, leaving businesses counting visits from customers who've never actually opened the website. Community managers have their own version of that problem. A lovely spike in engagement doesn't tell you much if you can't explain who's behind it.
What to watch next:
I'm still waiting to see what happens in Brazil. Discord's video restrictions hit the two-month mark on October 17, and the latest ANPD update leaves its appeal unresolved. I'd particularly like to know what evidence the regulator would consider sufficient to restore the service.
Next up is Gainsight Pulse Europe, October 19–20 in London. Gainsight has been developing AI moderation for direct messages and broader access to community knowledge through MCP. The Kyriba session on community engagement and adoption should be interesting, especially if we get some detail on the day-to-day support work involved.
Black Friday falls on November 27, too, and there’s still time to check who's handling moderation appeals and where unanswered community questions end up.
Preparing for the rush or dealing with platform changes? Get in touch. I'd love to hear how it's going.
What Did This Week Tell Us About Community Management?
The most pragmatic lesson from this week is that community teams can’t treat platform decisions as somebody else’s problem anymore.
Brazil's regulators aren't simply asking Discord whether safeguards exist. They want evidence that restrictions work. Reddit is moving moderators away from RSS toward supported Discord and Slack alerts while opening more explicit CRM connections for brands. Khoros is improving auditability and external knowledge access, and Gainsight is building community links into AI, support, and product workflows.
Discord might be grabbing the headlines, but I'd be asking some awkward questions about the systems behind these communities. Who's allowed to change moderation rules, and can anyone trace what happened afterward? If a platform makes a sudden change or a regulator wants answers, will the tools connected to customer support actually help the team figure things out?
FAQs
Is Discord banned in Brazil?
No. Discord itself remains available in Brazil. The current restriction applies to Go Live and equivalent video-sharing features. That distinction matters for anyone searching for a Discord Brazil ban update, because text chat, voice, servers, and other non-video functions continue to operate.
Why Does Reddit Care So Much About Marketing Spam?
Reddit works because people generally expect to hear from other people. That gets harder when brands coordinate accounts to recommend products while pretending they're ordinary users. As Reddit builds out its shopping and AI discovery features, it can't afford to have every enthusiastic recommendation look suspicious. There's money riding on those conversations being genuine.
Think spam detection, automated rule checks and help sorting through cases that need human attention. They're handy when moderators have hundreds of posts to review. But if an automated ban looks wrong, someone's still got to investigate it. Khoros's reporting changes and Discord's audit-log updates give teams more information to work with when members start asking questions.
When Is Gainsight Pulse Europe 2026?
Gainsight Pulse Europe runs October 19–20 in London. I'd be watching the Digital CS and AI sessions, particularly anything showing how community activity connects with customer health, retention and the work AI agents can take on.
