What the vendors say
On 23 September 2026 Zenity said Gartner named it a Market Shaper in the inaugural Emerging Market Quadrant for AI Application Security, Startup Vendors, one of two vendors in that category among nearly 20 assessed. The same day, Noma Security said Gartner placed it in the quadrant's Full-Spectrum, Context-Aware Agentic Security category as a Market Shaper.
On 15 September 2026 SentinelOne said Latio's 2026 AI Security Market Report named it an AI Security Platform Leader, and that its Prompt Security offering covers more than 15,000 AI applications and sites and extends to agents, MCP servers and custom-built applications.
On 28 August 2026 Harmonic Security summarized a Gartner Market Overview for AI Usage Control, which it quotes as predicting that by 2030, more than half of enterprises will use dedicated AI usage control technologies.
Each of these is a vendor's account of an analyst's work. We have not read the underlying reports, and we label every such claim vendor-stated.
What placements measure
Analyst placements assess a vendor's position in a market: vision, execution, product breadth, customers, momentum. The quadrant Zenity and Noma Security describe covers AI application security for startup vendors, which is broader than our question. A placement in it says the analyst found the vendor significant in that market. It does not say how specifically the product controls a single agent action on a laptop.
That is why analyst recognition appears only inside our public proof criterion, as one signal of what a buyer can check before a sales call, alongside documentation, attestations, backing and pricing. Zenity scores 8 there and Noma Security 8. Neither placement changes their scores on stopping an action before it runs, where Zenity scores 8 and Noma Security 7.
What the claims tell you about scope
The language in these announcements is useful in another way: it shows how each vendor sees its own scope. Zenity's announcement and its platform pages describe agents across SaaS, cloud, homegrown and coding environments, consistent with its tie for highest score on our cloud and workspace reach criterion. Noma Security's category name, full-spectrum agentic security, matches its coverage of endpoint, SaaS and homegrown agents. SentinelOne's figure of more than 15,000 AI applications and sites describes discovery breadth. Harmonic Security's choice of the AI usage control overview matches its focus on how people use AI tools.
How to use analyst claims in an evaluation
- Ask the vendor for the report itself, or the licensed excerpt, rather than relying on the press release.
- Check which market the report covers and whether it matches your question. A placement in AI application security is not the same as a placement in endpoint control.
- Check the date and the assessment criteria. Reports in a young category age quickly.
- Use placements to add vendors to a long list, not to remove them. Several products we score are too new or too narrow to appear in broad reports.
What we do instead
Our ranking answers one question from public product pages: which product controls what AI agents on devices and in cloud workspaces can do at the moment they act. Every score has a written reason and a source, and the method page shows the weights and the order on equal weights. Analyst placements and our scores answer different questions, and a buyer can use both.
A note on vendors not in the reports
Several products we score did not cite any analyst placement on the pages we read, including Bay, Lasso Security and Bloom Security. That says nothing about whether their products control agent actions well. Young vendors often appear in analyst reports a year or more after their products reach customers. Judge them on what their product pages describe and on what they show you in an evaluation.