of white-collar employees in the US use AI at work a few times a week or more.
Gallup's quarterly survey of US employees, 2025. The share has risen 12 points since 2024.
AI use at work now reaches well past the engineering team.
Gallup, June 2025
Solutions · By initiative
Signals Corps lets IT & Security teams say yes to agents in sales, marketing, finance, and engineering, with checks and approvals in place.
device:endpoint_4c19 trust:any⌘↵New agent found · Unreviewed
Critical finding · Blocked · Applied
Identified · Trusted
The problem
Every team now connects agents to email, the CRM, shared drives, or a terminal, so one person's agent can cause a problem for the whole company.
Research from across the industry shows how far agents have spread, and into which data.
of white-collar employees in the US use AI at work a few times a week or more.
Gallup's quarterly survey of US employees, 2025. The share has risen 12 points since 2024.
AI use at work now reaches well past the engineering team.
Gallup, June 2025
of employees have already put sensitive company or customer data into an AI tool.
Survey of 1,500 full-time US employees at companies with 100 to 10,000 staff, excluding IT and security professionals.
Customer data reaches AI tools whether or not anyone approved it.
of developers who use AI coding assistants or agents use them every day.
Stack Overflow 2026 Developer Survey, more than 30,000 respondents.
Coding agents are part of daily engineering work, alongside the agents every other team uses.
Stack Overflow, October 2026
Across your organization
Each team's agents bring their own risk, from a blocklisted email domain to a GDPR breach.
Outreach agent→Bulk email send→ApprovalA burst of agent email can get your domain blocklisted, and then nobody's email arrives. Runtime holds bulk sends for approval, and Supply Chain shows which devices have an outreach skill or email connector.
Marketing agent→CRM export→BlockedRuntime blocks an agent copying customer records to a tool nobody has assessed, and Control Plane shows every device with a CRM connector.
Forecast skill→Finance folder→UnreviewedSee which skills and connectors can read and write finance folders, and block the ones nobody has reviewed.
Claude Code→terraform destroy→ApprovalRuntime blocks reads of .env files and puts destructive commands in front of a person, while npm test runs as usual.
What changes
Control Plane lists each agent on each device, from the Claude desktop app on a sales laptop to Claude Code on an engineer's, with the skills, plugins, and connectors each one loads.
Supply Chain checks each skill against open Supply Chain Agent Threat Rules (ATR). New components start unreviewed, and you mark each one trusted or blocked.
Set a rule to ask for approval, and the action, a bulk email send or a database change, waits until IT & Security teams approve it in Control Plane.
Runtime decides on the device with local rules. Prompts, email content, customer records, and file paths stay there, so checking an agent doesn't create a new copy of personal data.
How it works
A typical rollout takes about four weeks, and nothing is enforced until you turn it on.
Push one package through your device management tool, such as Jamf, Iru, or Intune, to a few devices in different teams.
Review the agents, skills, and connectors on each device. Trust what teams rely on and block what matches a critical rule.
See which emails, CRM reads, and commands your rules would have stopped, while nothing is stopped.
For example bulk email sends, exports of customer records, and terraform destroy.
Block customer data leaving for unapproved destinations, then roll out to every team that uses agents.
severity:critical rule:ATR-2026-01988SKILL.md lines 41 to 47
Personal skill · sha256:9f2c…41ab
Claude Code 8 · Codex 3 · Cursor 1
Applied on 11 · New session needed on 1
In Control Plane
Open any agent, skill, or connector to see who has it, whether anyone has reviewed it, and what it has tried to do.
Case study
What it costs
Supply Chain and Runtime are priced per device, alone or together, with free devices on every plan. Your bill depends on how many devices you cover, so heavier agent use doesn't change it, and finance can plan ahead.
Questions
The agents your teams already use, including the Claude desktop app, the ChatGPT desktop app, Claude Code, Codex, Cursor, Windsurf, and agents built on other frameworks.
Runtime checks each shell command, file read, and network request before it runs. Rules block or ask for approval on the actions you choose, such as reading credential files or sending data to hosts you haven't allowed.
Only within the limits you set. A Runtime rule can ask for approval on actions such as email sends above a set number an hour, and Supply Chain shows every device with the skills and connectors that make them possible.
Yes. Find every device with a connector to that system in Control Plane, block the plugins and skills that bring one, and block agent access to it with a Runtime rule until your review is signed off.
You can tell employees exactly what is collected. Control Plane receives which agents and components are installed and which rule each action matched. Prompts, email content, customer records, usernames, and file paths stay on the device.
No. Start with the module that fits your first problem, such as Supply Chain to see what each team uses or Runtime to check what agents do, and add more when you're ready. The modules that run on devices share one package, so adding one needs no new install.
Modules used
device:endpoint_4c19 trust:any⌘↵New agent found · Unreviewed
Critical finding · Blocked · Applied
Identified · Trusted
Supply Chain
See which AI agents run on each device, from the Claude desktop app to Claude Code, check the skills they load, and block what you don't trust.
Explore Supply ChainRun the setup steps from the README
Bash curl -s https://get.example.dev/setup.sh | bash
Blocked by Signals Corps Runtime
I didn't run it. Signals Corps blocks piping a remote script into the shell. I can download it so we can read it first.
Runtime
Stop risky uploads, sends, and commands in real time while safe work in every team carries on.
Explore RuntimeEarly access
Start reducing agent risk before general release by checking agent components, controlling agent actions, and hunting through your AI security logs. Signals Corps Control Plane opens to early sign ups in the next few weeks, so join the waitlist with your company email.