Market · Weekly Roundup
The leading AI platforms are the public surfaces where OpenAI, Anthropic and Google show what their models can do, and in the week to 3 October 2026 the new pages leaned hardest on connection: a surge of Claude connectors, plugins and agent listings plus a new OpenAI business marketplace built the integration layer, while Gemini scaled a global library of image-generation styles and the labs stood up fresh trust, safety and compliance surfaces.
Based on source data gathered by reconnAI, covering new pages across the leading AI platforms in the week to 3 October 2026.
TL;DR
- The connector shelf kept filling: Anthropic poured dozens of new Claude connectors into its marketplace, from payments and finance tools such as Mollie, Cashfree, Revolut X, CoinMarketCap and D&B Hoovers to travel, events and data apps like GetYourGuide, All-Accor, Kahoot, Menti and Plotly Cloud, alongside new plugins such as OneSignal and Incident.io and an agent listing for Cyera, while OpenAI opened a business marketplace and partner pages.
- Image generation went global: Google scaled Gemini into a broad library of named image-generation styles, from arcade, bento and bloom to chibi, clay and elven, each shipped across many languages at once.
- Trust and safety surfaces scaled: Anthropic stood up a wave of safety, compliance and verification flows, and OpenAI published safety cases for frontier training, a piece on disrupting a coordinated model-distillation campaign, and Sign in with ChatGPT terms.
- Flagships and enterprise proof ran underneath: OpenAI DevDay 2026, GPT-6.1 Sol and named customer stories from Albertsons to Chatham Financial, plus Claude agents you can coach, a Frontier Academy and fresh customer and vertical pages.
Last week the centre of gravity was proof and enablement, as the majors surrounded the integration shelf with named customer stories and developer playbooks. This week the build kept going, but the emphasis shifted back to reach: Anthropic filled its marketplace with connectors and agents, OpenAI opened a business marketplace of its own, and Google turned Gemini image generation into a templated, global product. It follows straight on from proof and playbooks scaling as integrations climb, connectors surging as the majors court finance and law and Claude opening a plugin directory. Below is where the new pages clustered.
Where is the growth focus this week?
Net-new pages published this week, by theme
Bars are relative to the largest theme. Counts reflect new pages observed across the leading AI platforms in the week to 3 October 2026.
1. The connector and marketplace shelf keeps filling
The single largest cluster of new pages again built the connective tissue between the assistants and the software teams already run. Anthropic poured dozens of new Claude connectors into its marketplace across every category: payments and finance (Mollie, Cashfree, Revolut X, CoinMarketCap, financial datasets, D&B Hoovers, Onchain Diligence), marketing and sales (Constant Contact, Omnisend, Leadfeeder, Ubersuggest by Neil Patel, Loops, Pitch), travel and hospitality (GetYourGuide, All-Accor, Brek hotel wholesale booking, SlickTrip), events and engagement (Kahoot, Menti, RSVPify) and data and developer tooling (Plotly Cloud, CloudBees Unify, Informatica, Infobip documentation, SciSpace, USAFacts, Chatbase, Tana). It added fresh plugins such as OneSignal, Incident.io and an AWS Startup Advisor, and a new agents listing for Cyera. OpenAI answered by opening a business marketplace and partner pages, including a cloud partnership with MegazoneCloud. Watch it build on our Claude intel page and OpenAI intel page.
Why it matters: a horizontal assistant is hard to buy, but an assistant that already reaches the tools a team pays for is easy to adopt. Every connector, plugin and partner app puts the model inside an existing workflow, and each listing is a page a brand in that category can be named on. It extends the plugin shift we covered in ChatGPT retiring custom GPTs and the directory we tracked in Claude opening a plugin directory, now running at a larger scale. For a brand, being connectable is fast becoming table stakes for being reachable inside an answer.
2. Gemini turns image generation into a global style library
The clearest new emphasis this week came from Google. Gemini published a broad gallery of named image-generation styles, each with its own page and prompt recipe, running from arcade, bento and bloom to bronze, chibi, claw, clay, dappled and elven, alongside purpose-built looks such as a coach and a mural template. The striking part was the reach: each style shipped in many languages at once, including English, Spanish, French, German, Japanese, Korean, Portuguese, Arabic, Turkish, Dutch and Indonesian, so the same creative surface went live for a global audience on day one. It turns a single prompt box into a browsable, reusable catalogue of looks, the same templating instinct we saw in Gemini adding Skills.
Why it matters: naming and packaging a capability is how a platform makes it discoverable and repeatable. When a style has its own page, it becomes something people search for, link to and ask an assistant to apply, and the visual answer becomes as much a destination as the text one. For brands, it is a reminder that multimodal and visual results are a growing slice of how people get answers, a shift that also sits behind ChatGPT adding virtual try-on. Owning the named look, product or template in your category is becoming its own form of visibility.
3. Trust, safety and compliance surfaces scale
The third cluster built the governance scaffolding around all of the above. Anthropic stood up a wave of safety, compliance and verification pages, from a cyber use-case intake and a model-safety bug bounty to jailbreak and vulnerability reporting, UK and Australia online-safety reporting, content reporting, removal and appeal routes, and identity and life-sciences verification programmes, plus a data-retention request flow. OpenAI published its own trust surfaces: safety cases for frontier AI training, a piece on disrupting a coordinated model-distillation campaign, a proaction update, and the terms and policies for Sign in with ChatGPT. Read it against the safety week we covered in safety taking the front page as ecosystems widen and agents going to work across the enterprise.
Why it matters: trust is what lets an assistant be used for real, regulated work, and the labs are building it as an operational discipline rather than a policy footnote. Every reporting route, verification programme and published safety case tells a buyer, and a model, that the platform can be relied on. For a brand, the same logic applies in miniature: a visible data, security and compliance posture helps a model treat you as a safe choice when it assembles an answer inside a regulated field.
Flagships, agents and enterprise proof run underneath
Under the three big moves, several threads ran through the week. On models and developer platform, OpenAI held DevDay 2026 and published a recap and a practical guide to building on GPT-6, launched the GPT-6.1 Sol and Dots models and expanded Codex Originals, while Anthropic promoted Claude Opus 5.5 for long coding sessions and rolled out import memory. On enterprise proof, OpenAI added named stories spanning Albertsons reimagining retail, Chatham Financial, a Basis tax workbook built with Astra and a push to help small businesses put AI to work, while Anthropic shipped customer stories for IRC and Supermetrics. On agents and verticals, Anthropic published pieces on agents you can coach with Asana, giving companies more control over agents with NVIDIA and how its own sales team rebuilt inbound with managed agents, alongside a Frontier Academy programme and solution pages for financial advisors, small business, teachers and life sciences. Read it against the flagship field from GPT-6 Astra to Claude Opus 5.5, and the new surfaces in ChatGPT profiles and sites.
What it means for your AI visibility
If the labs are wiring their assistants into every business tool, packaging capabilities into named surfaces and wrapping the lot in trust, the brands that want to appear inside those answers should build in the same direction. A short checklist to work through this month:
- Make your product connectable: the fastest-growing surface remains the connector and plugin shelf, so expose a clear, callable interface and document how a team would wire yours in.
- Own a named surface: as platforms turn capabilities into browsable styles, templates and listings, give your own product, look or method a clear name and its own page a model can point to.
- Think beyond text: with visual and multimodal answers growing, make sure your images, demos and product views are public, labelled and easy for an assistant to surface.
- State your terms and governance: a visible data, security and compliance posture helps a model treat you as a safe choice inside regulated work.
- Shape offers by vertical: as the majors publish finance, retail, education and small-business pages, a sector-specific page for your own product helps a model place you in the right answer.
- Track the direction weekly: read the shifts against your own category with our intel dashboards and decide where to invest.
Not sure where you stand today? Start with is your brand ready for AI search, work through building a GEO strategy and how LLMs decide which brands to cite, then browse the full archive on our news page.
Market Intelligence
Want to know where the AI platforms are heading next?
We track what the leading AI assistants are building and how brands appear inside their answers, so you can plan around the shifts before they reach your category.
Talk to our team →What to watch next week
- Connector velocity: how fast the Claude marketplace and the new OpenAI business marketplace keep filling, and which categories get wired in next.
- Image surfaces: whether Gemini keeps expanding its style library and whether the other platforms answer with named visual surfaces of their own.
- Trust cadence: whether the run of safety, reporting and verification pages keeps pace with the pace of new capability.
- DevDay follow-through: whether the GPT-6.1 Sol launch and DevDay guides turn into named production deployments.
The takeaway for anyone competing for visibility inside these assistants: the platforms are racing to be more connectable, more discoverable and more trusted, not just more capable. Read which integrations, surfaces and safeguards each lab is prioritising and you will navigate the next few months far better. Start from the reconnAI home page or talk to our team to put it to work.
Frequently asked questions
What was the main direction of travel for the AI sites this week?
In the week to 3 October 2026, the leading AI platforms leaned hardest on connection and reach. Anthropic filled its marketplace with dozens of new Claude connectors, plugins and an agent listing, OpenAI opened a business marketplace of its own, and Google scaled Gemini image generation into a global library of named styles, all wrapped in a fresh wave of trust, safety and compliance surfaces.
Which new integrations stood out this week?
Anthropic added Claude connectors including Mollie, Cashfree, Revolut X, CoinMarketCap, D&B Hoovers, Constant Contact, Omnisend, GetYourGuide, All-Accor, Kahoot, Menti, Plotly Cloud, CloudBees Unify, Informatica and SciSpace, plus plugins such as OneSignal and Incident.io and an agent listing for Cyera. OpenAI opened a business marketplace and partner pages, including a MegazoneCloud cloud partnership.
What did Google ship for Gemini image generation?
Google published a broad library of named Gemini image-generation styles, from arcade, bento and bloom to chibi, clay and elven, each with its own page and prompt recipe. The styles shipped across many languages at once, including English, Spanish, French, German, Japanese, Korean, Portuguese, Arabic, Turkish, Dutch and Indonesian, turning image generation into a browsable, global catalogue.
How should brands respond this week?
Make your product connectable through a clear callable interface, give your own product, look or method a named and linkable page, make your visual and multimodal content public and labelled, state your terms and governance so a model can treat you as a safe choice, and shape offers by vertical so an assistant can place you in the right answer.
About reconnAI
reconnAI tracks how the major AI models represent topics and sources across ChatGPT, Claude, Gemini, Perplexity, Copilot and Google AI Overview, across multiple regions. We monitor how those models answer and how they change over time, so you can stay ahead of shifts in the AI landscape. Talk to our team to put this intelligence to work.
Based on source data gathered by reconnAI.