Daily Marketing News, September 30, 2026: Google may stop reviewing complaints from those who flood Business Profile support

Muse is Meta's family of AI agents that carry out tasks on a user's request. The company first launched personal agents for regular people, and now there's a separate set for small businesses. The main part of the set is connectors, meaning integrations with third party services. The new partners are Zoom, Figma, Canva, Stripe, GitHub, QuickBooks, Slack and Shopify, while Meta's own tools and analytics were connected earlier. Meta gives example scenarios: review your content and build a promo campaign, check cash flow and find places to save, draft replies to customers, evaluate a marketing plan. A mode where you give the agent business goals is in the works. According to Meta, the agent won't publish, send or spend anything without the owner's explicit approval, writes Social Media Today. For people who run Facebook and Instagram, this is a planning assistant, but you'll have to open up your business data to it.

The role appeared in Forum, a separate app for Facebook groups that pulls all your groups into one feed; it has been in testing since May and can only be downloaded in the US on iOS and Android. The new role replaces the old active member badges and the group expert role. It's given for useful contributions: answering questions, sharing personal experience, giving advice. The member gets a badge and perks such as faster post publishing. The admin assigns the role, and Meta also suggests candidates. The test is running in Forum in the US and in Facebook worldwide, Meta said. 2 more changes: Ask search in Forum now compares options and suggests new questions, and posts now have topic labels. For a brand that works through a community, this is a ready made reward for people who answer questions on its behalf.

The TikTok agency partner program has existed since 2020, and its directory lists more than 600 approved companies: creative agencies, marketing tech firms, ad measurement services. Previously all partners had the same status; now each agency will get one of 2 badges. Agency Partner goes to qualified agencies with strong results, operational capacity and a certified team. Premier Agency Partner is the top tier: a small group of the best partners worldwide will get it for proven business results and deep platform expertise. Both badges mean the agency has been verified against the requirements, writes Social Media Today. For advertisers, it's a ready filter when choosing a TikTok ad contractor. For agencies, it's a clear step to grow toward.

Threads is Meta's text based social network, and its live chats are real time group discussions; they live in the direct messages section, separate from the feed, and first appeared for the NBA playoffs. Before, you could only react to a single message. Now any participant can send an emoji to everyone at once, and it floats across the screen. The "raise hand" button lets you ask to speak, which matters in chats with a limited number of participants. Chats are coming to the web version of Threads soon, and the company says it's one of the most requested features. Chats have opened in Australia, and chats for "Dancing with the Stars" and the WNBA playoffs are launching together with creators, reported Social Media Today. For media and brands, it's a format for gathering people around an event while it's happening.

This is about the EU Digital Markets Act, or DMA, which forces the largest platforms to open up to rivals. The decisions were made in July: Google has to hand search data to other search engines and let competing AI services into Android. The company says it would have to give away private search history without sufficient anonymization and weaken Android's protections. According to Tech2Geek, the case was filed with the EU General Court on September 28, 2026. The same day Search Engine Roundtable published a clip from an interview with Nick Fox, Google's senior vice president of knowledge and information: he called the DMA driven changes the biggest drop in search quality in Google's 28 year history. For anyone doing SEO in Europe, this means search results and data access will keep shifting for a long time while the case goes on.

AI Overviews is the AI generated summary Google shows above regular search results. Next to the answer are cards for the sites it cites; on desktop they currently sit in the right column, and in the test they move below the answer text. According to Search Engine Roundtable, Google has tried similar versions at least twice before: in September 2025 and in March 2026. A third attempt in a year suggests the company is looking for a permanent place for these links, not just testing an idea. Whether someone notices a link to your site and clicks it depends on where these cards sit. Site owners who show up in these answers should track desktop clicks separately in Search Console to see if the test reaches them.

The local panel is a business card in search with buttons to call, get directions, visit the website and more. On mobile, Google is testing 2 rows of these buttons; earlier it tried a "+" button for extra actions and a scrollable carousel. The second change is a bug in Business Profiles and Maps: some businesses' pins sit in the middle of the ocean. Reports suggest it affects service area businesses that don't show their address publicly: mobile technicians, delivery services, in home services, writes Search Engine Roundtable. If you manage a listing like that, open it in Maps today and check where the pin is: someone who sees your business in the water is unlikely to call.

According to TechCrunch, the model came out on September 28, 2026. It's the second model in the Claude 5.5 family after the top tier Opus 5.5, which was unveiled on September 22, 2026; the previous Sonnet 5 came out about 3 months ago. Anthropic says the new model responds more than 30% faster and uses fewer tokens, so a task costs less, even though the API price hasn't changed. The model is available on Amazon Web Services, Google Cloud and Microsoft Azure, and the smaller Haiku 5.5 will follow later. On the FrontierCode programming benchmark, Sonnet 5.5 scores 52.1% and the top tier Opus 5.5 scores 54.4%, a gap of 2.3 points. If you use AI to write copy and reports or to automate routine work, this is nearly the top model for less money.