Market · Weekly Roundup
In the week to 25 July 2026, Anthropic filled the Claude plugin shelf with around twenty developer and security integrations while OpenAI widened its partner network and launched OpenAI Presence and a ChatGPT small-business program - the direction of travel is the integration layer, with agent security wiring in underneath.
Based on source data gathered by reconnAI, covering new pages across the leading AI platforms in the week to 25 July 2026.
TL;DR
- The plugin shelf fills up: Claude added around twenty new plugins in the window - data and DevOps tools like Azure SQL, MLflow, GitKraken, New Relic, GrowthBook and Qodo, plus a ServiceNow connector - turning Claude Code into a hub that reaches into the tools teams already run.
- OpenAI widens the channel: a fresh wave of named delivery partners went up - Fujitsu, Globant, ZS, Tredence, Snorkel AI, CloudWERX, Blend360 and more - alongside the launch of OpenAI Presence and a ChatGPT small-business program.
- Agent security wires in: a CISO guide to agentic AI, security plugins (CrowdSec, StackHawk, a Claude security pack, DeepEval), a write-up on how Anthropic secures its AI-native SDLC, and OpenAI work on long-horizon safety-alignment.
- Proof and enablement keep pace: role-based Claude certifications, a model-selection guide, customer stories (Caylent, Eve Legal, the Epilepsy Foundation), and OpenAI pages on health in ChatGPT, national science and board appointments.
The models kept shipping, but the week to 25 July 2026 was about the layer around them. The clearest signal was a rush of pages devoted to plugins, partners and connectors - the plumbing that lets a capable model plug into the tools a business already runs. It carries straight on from last week’s OpenAI signs up the consultancies, and from the enterprise tool stack and ecosystems and security roundups before it. 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 25 July 2026.
1. Anthropic fills the plugin shelf
The biggest single push of the week was a run of plugin pages on Claude - around twenty of them, each wiring a named third-party tool into Claude Code. The list reads like a working engineer’s toolbelt: data and analytics tools (Azure SQL Developer, MLflow, GrowthBook, Altimate Code), developer workflow (GitKraken, Qodo, Browser Use, CKEditor, a UI theme designer and modern web guidance), observability (New Relic), and work-management ties (monday.com), with a fresh ServiceNow connector alongside. It is the same marketplace instinct we traced in the connectors and integrations and verticals and marketplace weeks.
These are not model announcements - they are reach. Every plugin is one more place Claude can do real work without leaving the tool a team already lives in, and you can watch the shelf grow on our Claude intel page. It is the productised version of the move that took Claude into Microsoft 365 Copilot, and it sits on top of the model line from Claude Sonnet 5 and Claude Opus 4.8.
Why it matters: a plugin shelf is a distribution strategy. When the tools a company already trusts can call a model in place, that model stops competing on benchmarks and starts competing on how many jobs it can quietly complete. For brands, the lesson is the same one we set out in how LLMs decide which brands to cite - be the source an assistant can reach when it acts, not just the page it reads.
2. OpenAI widens the partner network
OpenAI answered from the channel side, adding a fresh wave of named delivery partners - Fujitsu, Globant, ZS, Tredence, Snorkel AI, Merantix Momentum, CloudWERX, Blend360, Rosetree Solutions and a long tail of specialists. It extends the wall of consultancies and integrators we covered in OpenAI signs up the consultancies, and you can see the picture build on our OpenAI intel page.
Two launches sat on top of the partner run. OpenAI Presence got its own business surface and introduction, and a ChatGPT small-business program opened with a dedicated lead page - a deliberate reach past the enterprise into the long tail of smaller buyers. A new reserved-tier API page rounded out the commercial plumbing.
Why it matters: partners win the rollout and a small-business program widens the funnel. Together they say OpenAI wants to be bought at every size of company, not just the Fortune 500 - the same land-everywhere pattern that took ChatGPT into work and, this month, health.
3. Agent security wires into the stack
The theme threading through both platforms was security and governance for agents. Anthropic published a CISO guide to agentic AI, a write-up on how it secures its own AI-native software development lifecycle, and a story on how Outtake built a cyber investigator on Claude - and several of the new plugins were security tools in their own right, from CrowdSec and a pair of StackHawk integrations to a Claude security pack and the evaluation tool DeepEval. OpenAI matched the direction with research on safety-alignment for long-horizon models, a scorecard for the AI age, and a candid model-evaluation security disclosure.
This is the governance conversation that now travels with every agent that can act on real systems, and it echoes the moment OpenAI opened a security front. It keeps pace, too, with the guardrails story we tracked in Google’s llms.txt clarification.
Why it matters: as agents move from answering to doing, security stops being a footnote and becomes the thing a board signs off. The platforms publishing CISO guides and secure-SDLC pages are arming the exact people who approve - or block - an enterprise rollout.
Proof and enablement underneath
Beneath the three big moves, the enablement layer thickened. Anthropic shipped role-based Claude certifications, a guide to choosing the right model for a use case, and a set on verification loops and voice-mode reasoning, plus fresh customer stories from Caylent, the legal platform Eve, the Epilepsy Foundation and the National Domestic Workers Alliance. A Datadog build and new working at the frontier pages for Cursor and Rakuten added named proof. OpenAI filled in around its channel with board appointments, a page on how news organisations use AI, national-science and AI-infrastructure pieces, and a creative Codex collaborator story.
It is the same proof-and-teach pattern we flagged from wire up, skill up and prove it onward - certifications teach the channel, customer stories reassure the buyer, and model-selection guides lower the bar to getting started.
What it means for your AI visibility
If the labs are building outward into plugins, partners and security, the brands that want to appear inside those answers should build in the same direction. A short checklist to work through this month:
- Be reachable, not just readable: as models act through plugins and connectors, make sure your product and data can be called in place, not only found in a search box.
- Map to the channel: if your buyers work with the integrators now certified on these platforms, make that relationship explicit on your public pages so a model can place you.
- Publish your proof: named customer stories and clear security and governance pages are what carry a brand - and an assistant - through procurement.
- Lead with governance: as agent security becomes a board-level question, a visible, well-structured security posture helps you get cited in the answers buyers trust.
- Write for tasks, not just queries: being the source an agent uses to finish a job now matters as much as ranking for a question.
- Track the direction weekly: read the shifts against your own category using 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, 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
- Plugin velocity: whether Claude keeps adding integrations at this pace, and how quickly OpenAI answers with a marketplace of its own.
- Partner spread: which regions and specialisms fill in next on OpenAI’s partner directory, and whether Anthropic stands up a formal one.
- Security depth: how far the CISO and secure-SDLC material goes, and whether governance becomes its own named surface.
- Small-business pull: how hard OpenAI pushes the ChatGPT small-business program, and whether the labs localise it beyond English.
The takeaway for anyone competing for visibility inside these assistants: the platforms are racing to become the place work gets connected, delivered and secured - carried by plugins, backed by partners and proof, and governed for the boardroom - not just the place questions get answered. Read which integrations, partners 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 25 July 2026, the new pages across OpenAI and Anthropic leaned into the integration layer. Claude added around twenty new plugins and a ServiceNow connector, OpenAI widened its partner network and launched OpenAI Presence and a ChatGPT small-business program, and both platforms pushed agent security and governance content. The models advanced, but the centre of gravity sat in the ecosystem around them.
What went into the Claude plugin shelf this week?
Around twenty plugins landed, spanning data and analytics (Azure SQL Developer, MLflow, GrowthBook), developer workflow (GitKraken, Qodo, Browser Use, CKEditor), observability (New Relic), work management (monday.com) and security (CrowdSec, StackHawk, a Claude security pack, DeepEval), with a new ServiceNow connector alongside.
What did OpenAI Presence and the small-business program add?
OpenAI Presence gained a dedicated business surface and an introduction page, while a ChatGPT small-business program opened with its own lead page - a deliberate reach past the enterprise into smaller buyers. A new reserved-tier API page rounded out the commercial plumbing, alongside a fresh wave of named delivery partners.
How should brands respond to a rising agent-security layer?
Make your product reachable through plugins and connectors, map to the integrators the labs are certifying, publish named customer and governance proof, and lead with a clear security posture. Our guide to how LLMs decide which brands to cite walks through the detail.
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.