Snapshot
Continue to About ↓- What it does
- Opsin is an enterprise AI security and governance platform that connects to sanctioned AI platforms such as Microsoft Copilot, Copilot Studio, ChatGPT Enterprise, Claude Enterprise and Google Gemini to inventory employee-built agents and map each agent's intended purpose against the permissions, data and tools it can actually reach.
More detail
- It helps security teams prioritize and remediate the riskiest agents, including runtime blocking or quarantine, and addresses sensitive data oversharing in generative AI results.
- Best for
- Mid-to-large enterprises rolling out Copilot, ChatGPT Enterprise, Claude Enterprise or Gemini, where security and IT teams need an inventory of employee-built agents and a way to triage over-permissioned or abandoned ones.
- Pricing
- Opsin's pricing model is not disclosed.
More detail
- The site offers a free assessment and a demo request, but no published price, tiers, minimums or contract terms.
- As a cloud SaaS product sold to enterprises, expect a sales-negotiated agreement; fees, minimums and any implementation costs are unknown.
- Main limitation
- Individual consumers or teams without staff capacity to act on findings — a public reviewer noted the tool is simple but implementation requires someone dedicated to addressing open issues.
More detail
- Platform coverage is limited to supported AI platforms rather than every AI tool or homegrown LLM app.
- Record
- Living and monitored
- Reviewed
- Evidence
- Public web research
- Machine access
- Structured profile ↗
About
- Opsin is an enterprise AI security and governance platform that connects to sanctioned AI platforms such as Microsoft Copilot, Copilot Studio, ChatGPT Enterprise, Claude Enterprise and Google Gemini to inventory employee-built agents and map each agent's intended purpose against the permissions, data and tools it can actually reach.
- It helps security teams prioritize and remediate the riskiest agents, including runtime blocking or quarantine, and addresses sensitive data oversharing in generative AI results.
What it helps with
- Unknown and abandoned AI agents created by employees across sanctioned AI platforms, including draft agents never publis
- Over-permissioned agents and sensitive data oversharing in generative AI results
- Difficulty telling whether an agent's connection to sensitive systems is legitimate or risky
- Risk triage without context on what each agent was built to do versus what it can reach
- Slow investigations into agent-related exposure
Capabilities
A practical view of what is available and where limitations apply.
| Offering | Available | Detail |
|---|---|---|
| AI Agent Discovery | Yes | Continuously inventories employee-built agents across supported AI platforms, including drafts, with owner, last activity, permissions, tools and data reach. |
| Ownership resolution | Yes | Identifies who actually built each agent rather than relying on a default owner assignment, so remediation routes to someone who can act. |
| Intent-to-capability mapping | Yes | Classifies what each agent was built to do and compares it with the data, permissions and tools it can actually reach. |
| Risk prioritization | Yes | Ranks agents by access, data and capabilities, with plain-language reasoning behind each rating. |
| Runtime monitoring and containment | Yes | Compares live activity to intent and lets teams block or quarantine a risky agent, recheck the issue and notify its owner. |
| Oversharing protection | Yes | Detects and helps remediate sensitive data exposure in AI results, including legacy permission sprawl; customer outcome cited by Opsin: Culligan cut sensitive data in Copilot results from 80% to under 15%. |
| Agentless deployment | Yes | Uses delegated API access with no endpoint software, browser extensions or employee surveys. |
| Public developer documentation | Unknown | No public documentation site, API reference or knowledge base appeared in the evidence pack. |
| Self-serve or consumer offering | No | No self-serve or consumer offering is evidenced; the site offers a demo and free assessment. |
Bot trust
How confidently AI systems can verify this profile from public evidence.
—of 100
ChatGPT—/100
Claude—/100
This profile has not completed a Bot Score assessment yet.
Evidence
Where this record came from, when it was reviewed, and how confidently it can be used.
- Verification
- Scribble researched these facts from the linked public sources.
- The owner has not manually confirmed every field.
- Last reviewed
- Primary source
- opsinsecurity.com
Technical profile details
Record ID
68ac9d87-826e-449c-999d-a3211d664aa8Canonical URLhttps://profiles.scribble.network/listing/opsin-68ac9d87