Industries · Protective Services · Record Substrate
An agent accessed criminal-justice data. A CJIS auditor will ask what it did.
CJIS gives AI no exemption: an agent touching criminal-justice information inherits the same auditable-record controls as any system. Built by a veteran-led team, the substrate produces the record of who accessed what and which model answered, inside your own boundary, designed to run air-gapped.
§ 01 The problem
Agents are touching CJI. The log is not keeping up.
The moment AI agents retrieve and act on criminal-justice information, the trail of who accessed what, and which model produced the answer, is easy to lose. That is not only a technology problem. It is an accountability problem under a policy that expects the log.
Who accessed that CJI, what came back, and which model answered? Agent handoffs and model calls leave no reviewable record by default. Retrieval history lives inside vendor APIs, so when a model changes, the record changes with it, if it survives at all.
That is an operator problem. Subnet345 is built by a veteran-led team that runs a governed multi-agent operation every day.
§ 02 Why now
Can you show what your agents did with CJI?
AI agents moved from controlled pilots to production faster than any previous public-safety technology. The governance acceptable in a pilot is not acceptable against criminal-justice information. CJIS v6.1 raises the floor at the same time.
2023 to 2024
Pilots. Agents away from CJI.
Single models in sandboxes, isolated from criminal-justice data. The audit question had not arrived because agents were not yet touching CJI.
2025 to 2026 · Now
Production. Agents touch CJI.
Agents assist with records, case work, and retrieval over criminal-justice information. CJIS gives them no exemption: the access must be logged, attributed, and kept inside agency control, today.
2026 to 2027
Enforcement. The auditor asks.
CJIS v6.1 phases into full compliance by 30 September 2027, mapped to the NIST 800-53 moderate baseline. The FBI's most common audit findings are inadequate audit logging and encryption gaps, which is exactly where AI tooling adds new risk.
The audit question
If a CJIS auditor asked you to show every access to criminal-justice information your agents made, and which model produced each answer, could you?
§ 03 Use cases
Where the substrate lands first.
An attributable record is the precondition for letting agents touch criminal-justice information at all. The agencies that build governed AI agents first carry an advantage into every audit that follows.
Records and case work on CJI
Agents that assist with records, reports, and case files touching criminal-justice information, with access attributed and reviewable. Speed on the work without losing the log an audit expects.
Retrieval with the model on the record
When an agent retrieves CJI to answer, the record captures who asked, what was returned, and the model and version that produced the output. The answer is accountable, not anonymous.
Data stays inside agency control
Coordination, policy, and the record run inside your boundary, and the substrate is designed to run air-gapped. The one path that reaches outside by default is the model an agent calls; bringing the model in-boundary too, so criminal-justice data never leaves at all, is the air-gapped path we build toward.
§ 04 Against the policy
Measured against what CJIS expects.
The CJIS Security Policy, mapped to the NIST 800-53 moderate baseline, expects access to criminal-justice information to be attributed, logged, retained, and kept under agency control, with no exemption for AI. This material does not constitute legal or regulatory advice, and CJIS compliance is a property of your agency and its personnel, not of the software. Assess your obligations with your CJIS Systems Officer.
Sources: FBI CJIS Security Policy v6.1, 25 June 2026 (CJIS Security Policy Resource Center); NIST SP 800-53 (Rev 5, moderate baseline). Regulatory mapping reviewed: August 2026. Informational, not legal advice.
§ 05 What we ship
The substrate, run by the team that operates one every day.
Subnet345 builds and operates the record substrate for AI agents. We do not advise on governance from the outside. We run an audited multi-agent operation ourselves, inside our own perimeter, on the same substrate we would stand up in yours.
01 / Preserve
Access to CJI
02 / Reconstruct
Who, what, which model
03 / Avoid lock-in
Keep governance portable
Result
Audit-ready always
Built by a veteran-led team
The founding team includes U.S. Navy and U.S. Army veterans and operators who ran the infrastructure beneath national-scale security systems. We build for this environment because we came from it.
Immerse before we onboard
We stand the platform up inside your environment first, with your operators, against your telemetry, before you commit to a full rollout. We prove it before we scale it.
Transfer is the deliverable
You should be able to operate the platform without us at any point and keep running. We prove that condition before we step back, not in a proposal.
§ 06 Onboarding
How onboarding starts.
Every rollout begins with a structured conversation and ends with a transfer. No open-ended dependency, no shadow team after the sale. The engineer who stands the platform up is the one who trains your team.
Initial conversation · 45 minutes
Understand your current AI agent program, where it touches criminal-justice information, and whether this is the right fit for both organizations.
Governance assessment · Structured
A structured evaluation of your AI deployments against CJIS audit, access, and retention requirements, run inside your environment.
Design partner program
Early criminal-justice agencies become design partners, shaping the platform alongside the team that operates one every day.
Related
Adjacent sectors, and the substrate itself.
The question we help you answer
Can you prove which agent touched criminal-justice data, and which model answered?
If you cannot answer that today, and a CJIS auditor asks it tomorrow, Subnet345 is the substrate that changes that answer, inside your own boundary.