AIUC-1 certifies. AgentStatus verifies, every day.

Why we're reaching out

Certification is point-in-time. Reliability is continuous.

AIUC-1 is the first AI agent standard with real teeth, six pillars, quarterly updates, MITRE as a technical contributor, Schellman as the independent auditor. UiPath's certification required 2,000+ technical evaluations at audit time.

The Reliability pillar, agents behaving predictably and consistently in production, is exactly what we measure continuously. Not at certification. Every day. From outside the customer's stack, on the same networks their users are on.

That's the gap we'd like to fit into: between the audit and the renewal, generating the evidence that the bar is still being cleared.

What AgentStatus is

We continuously test AI agents and check answers with gold/contract and consistency probes (scoped — stably wrong but consistent still needs a domain expert).

Scheduled tests run from an independent network of 2,500+ consumer devices in 70 countries, with expected answers and drift detection when behaviour slips. Repeatable proof from residential observer networks — inbound reach past CDN/WAF — not from two AWS regions.

That includes multi-turn conversations and tool handoffs, so you can show what was exercised, from where, and what changed week over week.

Where we fit

Certification and continuous user-side evidence are different jobs.

Your platform

AIUC, Their lane

• • <strong>Defines the standard</strong> (AIUC-1, six pillars)
• • <strong>Audit + certificate</strong> via Schellman
• • <strong>Insurance backstop</strong> when things go wrong

Outcome

System of record

Dashboards, exports, lifecycle tools, and orchestration remain yours. We do not replace that surface.

AgentStatus

AgentStatus, Our lane

• • <strong>Generates the evidence</strong>, continuously
• • <strong>User-side tests</strong> between audit dates
• • <strong>Prevention signal</strong>: catches drift before it becomes a claim

Outcome

User-side layer

Reachability (Monitoring / Reliability) from residential networks, then outcome verification once reached — Scenario, Compositional, Safety, Stability; Consistency and Drift scores. Not prose matching.

One sentence. Certification answers "did we meet the bar then?" AgentStatus answers "is it still true this week, from real places on the internet?"

Proof of scale

Here is what we have already run, with plain definitions.

~10M test runs in ~2 months across the network. 8,000+ agents being tracked, including ones from companies you'd recognise (specifics under NDA).

We've also caught node operators trying to game the network with datacenter VMs instead of real consumer devices, the same kind of adversarial behaviour AIUC-1 is designed to make harder. Detection is built into the product.

What we are not claiming

We are an independent layer that runs alongside your stack.

We are not AIUC-1 auditors, not AIUC-1, and not an insurance company. We do not replace AIUC's standard or policies. We're the continuous evidence layer that sits between them.

What we'd like from this conversation

We would like to start with a two-week sandbox pilot.