Market · Weekly Roundup
In the week to 20 June 2026, the big AI platforms shipped far more plumbing than headlines. Across Claude, OpenAI and Perplexity, the new pages point in one direction: wiring the assistants into the tools companies already use, packaging up agents and skills, and stacking up proof that real customers are buying in.
Models still grab the attention, but the build-out this week happened around them. The clearest signal is not a new frontier model — it is the sheer volume of integrations, agent tooling and customer evidence going live. Below is where the week’s activity actually clustered, and what it tells you about each lab’s priorities.
The three things the labs built 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 20 June 2026.
1. Integrations are the new battleground
By far the largest push of the week was connective tissue. Claude added hundreds of new connector and plugin pages — entries for tools spanning CRMs, data warehouses, design suites, finance platforms and developer services. The message is blunt: the assistant is only as useful as the systems it can reach, and the race now is to reach everything.
Perplexity pulled in the same direction from a different angle, putting its new “Computer” agent inside the tools people already live in — Microsoft Teams, Slack, and Word, Excel, PowerPoint and Outlook — alongside connectors for platforms such as Snowflake. The pattern across both is consistent: meet the user inside their existing workflow rather than asking them to come to a chat box.
Why it matters: when an assistant can act across your real stack, the choice of model becomes a choice of ecosystem. The platform with the deepest, most trusted set of integrations starts to look less like a chatbot and more like a control layer over the whole toolset.
2. From chat to agents and skills
The second-largest cluster was enablement: a wave of new tutorials, guides and product pages teaching people to build with agents and skills rather than just prompt a model. New material covered building multi-agent systems, equipping agents with reusable skills, agent SDKs, and a fresh agentic workspace product (“Cowork”) aimed at teams.
Perplexity’s “Computer” is the same idea wearing different clothes — an agent that does work on your machine and inside your apps, now reaching everyday Mac and Windows users. OpenAI nudged in the same lane with new pages on deployment simulation and enterprise controls. The throughline: the unit of value is shifting from a single answer to a unit of work an agent can carry out end to end.
Why it matters: labs are no longer selling intelligence by the message. They are teaching customers to wire agents into processes — which is far stickier, and far harder for a rival to displace once it is embedded.
3. Proof, verticals and the enterprise land-grab
The third cluster was evidence. More than two hundred new customer case-study pages went live alongside industry solution pages for financial services, healthcare, life sciences, the legal industry, government, education and nonprofits. This is the language of a sales motion that has matured past “try it” into “here is who already runs on it, in your sector.”
OpenAI leaned hard into one vertical in particular — health and science — with new pages on improving health intelligence in ChatGPT, work on diagnosing rare childhood diseases, and a life-sciences benchmark, plus enterprise spend controls and a new partner network. Perplexity, meanwhile, sharpened its enterprise story around finance and data science. Different sectors, same instinct: pick the high-value industries and build dedicated proof for each.
Why it matters: verticalised proof is how these platforms move from individual seats to enterprise-wide contracts. Case studies and industry pages are the quiet infrastructure of that shift — and there were a lot of them this week.
The quiet one: Gemini
Not every platform moved. Google’s Gemini surface was effectively static this week — no meaningful new pages in the public-facing properties we follow. That does not mean nothing is happening internally, but on the evidence of what shipped, the visible momentum belonged to Claude, Perplexity and OpenAI.
What to watch next week
- Integration velocity: does the connector and plugin count keep climbing, and which categories (finance, design, data) fill in fastest?
- Agent products going GA: whether agentic workspaces like Cowork and Perplexity’s Computer move from preview to general availability.
- Vertical depth: which industries get the next round of dedicated solution and case-study pages — healthcare and financial services are clearly in front.
- Gemini waking up: any sign of Google answering the integration and agent push with new public surfaces of its own.
The takeaway for anyone competing for visibility inside these assistants: the platforms are racing to become the place work gets done, not just the place questions get answered. The brands that understand which integrations, agents and industries each lab is prioritising will read the next few months far better than those still watching only the model announcements.
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 →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.
Based on source data gathered by reconnAI.