is the minimum time organizations using a high-risk AI system must keep the logs it produces.
EU AI Act, Article 26(6), for deployers. Longer periods can apply under other EU or national law.
You need logs you can keep, find, and produce on request.
Solutions · By initiative
Signals Corps keeps the record regulators and auditors ask for: which AI agents you run, what they can reach, and what they did.
Run 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.
The problem
The EU AI Act, GDPR, and audit frameworks all expect you to show what your AI does, and agents leave few records.
Under the EU AI Act, any organization using a high-risk AI system, such as one that screens job applicants or assesses credit, must keep the logs that system produces for at least six months. GDPR asks where personal data goes. ISO/IEC 42001, SOC 2 auditors, and customer security reviews ask which AI tools you use and what controls you have. Agents running on employee devices rarely leave a record anyone can show.
The EU AI Act sets the dates and the stakes.
is the minimum time organizations using a high-risk AI system must keep the logs it produces.
EU AI Act, Article 26(6), for deployers. Longer periods can apply under other EU or national law.
You need logs you can keep, find, and produce on request.
is when high-risk obligations apply to AI used in areas such as hiring, credit, and education.
2 December 2027 for Annex III uses, moved from August 2026 by the AI Omnibus, published in July 2026.
There's time to put records in place before the rules apply.
Future of Privacy Forum, 2026
or 3% of worldwide annual turnover, whichever is higher, is the maximum fine for breaching these obligations.
EU AI Act, Article 99(4), covering the obligations of providers and deployers, including Article 26.
A missing record can cost more than keeping one.
What changes
Supply Chain lists every agent, skill, plugin, and connector on every device, with versions and trust decisions, so the AI register answers itself.
Runtime logs each action a rule matched, the rule, and the outcome: allowed, approved, or blocked, with who approved it.
Set how long Control Plane keeps the decision log, from six months to the period your sector requires, and export it for an auditor.
Prompts, email content, customer records, and file paths stay on the device, so the record itself doesn't create a GDPR problem.
How it works
Push one package through your device management tool and get every agent and connector on every device.
Turn on rules for customer data, credentials, and production changes, and set how long to keep the log.
Export the inventory and the decision log for an auditor, a regulator, or a customer security review.
Example deployment
Illustrative example
This company is fictional. The numbers come from a modelled fleet, not a customer.
Read the exampleWhat 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
One set of records covers you either way. The Act treats agent use as high-risk in areas such as hiring, credit, and education, and most coding and office work falls outside them. Many companies keep the same records for all agents, because auditors and customers ask for them too.
For high-risk uses listed in Annex III, from 2 December 2027, after the AI Omnibus moved the date from August 2026. For high-risk AI built into regulated products, from 2 August 2028.
The log is built to stay small. It holds which agent acted, the rule it matched, and the outcome. Prompts, email content, customer records, usernames, and file paths stay on the device and never reach Control Plane.
Signals Corps gives you the records that EU AI Act, GDPR, ISO/IEC 42001, and SOC 2 readiness all rest on. Supply Chain keeps your AI register current on every device. Runtime keeps a decision log for as long as you need it and enforces your AI policy with rules that allow, ask for approval, or block. You export the evidence when an auditor, a regulator, or a customer asks. You'll need all of this, because each framework expects you to show which AI you run and what controls you have, and agents leave no record of their own. Your legal team still decides which uses count as high-risk and when you're compliant, and this page isn't legal advice.
An export of every agent and connector on every device, the trust decisions on each, and the decision log for the period they ask about.
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 and which identities they can reach, 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 unsafe agent actions in real time while safe work 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.