Platforms

What keeps running after we leave.

Five products your team runs on its own. Each is licensed per module, activated by contract, and delivered as configuration and connectors, never custom code.

Platform security and compliance
$150/mo per module

AI Compass

Where the company's AI governance lives.

RACI, policies, guardrails, ethics, and the committee gate in one place, with version and owner. The client maintains it and runs their own CoE without Criterium present.

  • Live record per initiative, with owner and previous gate
  • Monthly gate: continue, adjust, or stop
  • AI maturity visible to the committee
compass.criterium-ai.com / portfolioSample data
AI maturity3.4
Live initiatives12
With owner12/12
Collections triageContinue
Intercompany reconciliationAdjust
Commercial proposal assemblyContinue
Monthly gatecontinue · adjust · stop
$150/mo per module

AI Agent Factory

The approved case becomes a packaged agent.

An agent builder for cases that already cleared the gate. Tools, guardrails, evaluation dataset, and oversight threshold are declared in the agent, ready for your team to publish in your own environment.

  • Tool schemas against SAP, Salesforce, or email
  • Evaluation against baseline before publishing
  • Human oversight above a declared threshold
app.criterium-ai.com / agent-factorySample data
ToolsSAP · Salesforce
Guardrails5
Evaluation dataset120 cases
Human oversightamounts > $50k
Evaluation against baseline94%
Ready to publish
$150/mo per module

AI Academy

Training on material from your own engagement.

The LMS each cohort runs in: content built from the engagement, per-session assessment, certification, and a progress board for the sponsor. The content stays as a company asset.

  • Learn, assess, and certify
  • Per seat or per cohort
  • Progress board for the sponsor
Open AI Academy
academy.criterium-ai.com / cohortsSample data
AI literacy5/8 sessions
Executive AI strategy3/8 sessions
Assessment and certification per session
$150/mo per module

ARIA MCP

The agent that works where your people already work.

ARIA operates connected to your sources through the Model Context Protocol: it reads, queries, and proposes actions in the operating systems. Write tools flagged as sensitive require human approval, and answers cite their sources where available.

  • Connectors with declared scope per tool
  • Human approval on sensitive tools
  • Audit log per call
aria.criterium-ai.com / mcpSample data
Which files have broken payment promises and active collateral?
File 18-402 has three broken payment promises and active collateral. Suggested route: restructure over 12 months, subject to human approval.
Sources · CRM · contract 18-402 · collections log
Write held for approval
$150/mo per module

ARIA Graph

What your operation knows, in a graph you own.

The graph holds the operating memory: which customer, which contract, which collateral, what was promised and when. ARIA reasons over it and traces the path of every answer back to its evidence.

  • Entities, relationships, and evidence
  • Traceability for every answer
  • Owned by the client
aria.criterium-ai.com / graphSample data
CustomerContractCollectionsCollateralPaymentsAudit log

Which product fits your operation?

Five products your team runs on its own. Each is licensed per module, activated by contract, and delivered as configuration and connectors, never custom code.