AgentStatus × Ema, a quick map of how we fit
Independent verification for Ema's AI employees.
We do two jobs: reachability from residential networks (past CDN/WAF), then reliability once reached — gold/contract and consistency checks, across the channels each platform supports, from 2,500+ nodes across 70 countries. We sit alongside Ema's Generative Workflow Engine, EmaFusion model, and pre-built AI employee library. We don't replace them.
What we understand about Ema
Ema is a horizontal agentic OS built for enterprise deployment.
Ema is the horizontal agentic OS where enterprises conversationally build AI employees that take on roles across the organization, claim validation, agent QA, compliance analyst, ticket resolution, proposal writing, prior authorization, and dozens more. The platform is powered by Ema's proprietary Generative Workflow Engine™ for multi-agent orchestration and the EmaFusion™ model that combines 30+ specialized models into one accuracy-tuned system.
For enterprise deployment, Ema operates both on-cloud and on-premise, with compliance across SOC 2 Type I & II, HIPAA, GDPR, ISO 27001, NIST CSF, NIST SP 800-171, NIST AI RMF, and ISO 42001, the world's first AI management system standard. Customers include Envoy Global, TrueLayer, and Moneyview, with deployments live across 200K+ employees.
What AgentStatus is
We measure whether users can reach the agent, then whether it still passes its checks.
Reachability. Controlled validations from 2,500+ residential devices across 70 countries measure whether users can open the agent the way they do — past CDN, WAF, and bot walls. Multi-geo is observer vantage for access and last-mile latency — not answer localization by probe IP, and not agent tool egress.
Outcome verification. Once reachable, we verify outcomes: Scenario (did it finish the job?), Compositional (do the pieces hold together?), Safety (must-not-say / policy / attacks), Stability (same ask, same story?). Consistency and Drift track what changed. Job anchors and side-effects where they exist — not prose matching. Optional sample review is corroboration only; stably wrong still needs a domain expert.
That includes multi-turn conversations and multi-agent journeys when customer paths span tools, escalations, and handoffs. It supports governance and risk conversations when stakeholders ask what was tested, from where, and what changed.
User-side validation is two separate jobs
Reachability
Outcome
Can we talk to it?
Residential observers take the inbound path customers take — past CDN, WAF, and bot walls that treat datacenter synthetics differently. Monitoring asks: is it healthy right now? Reliability asks: does it keep working over time? “Up” means reachable from home networks, not from AWS.
Outcome verification
Outcome
Did it do the right thing?
Reachable and wrong is still broken. Scenario — did it finish the job? Compositional — do the pieces hold together? Safety — must-not-say, policy, attack probes. Stability — same ask, same story? Scores: Consistency and Drift. Not prose matching. Optional sample review is corroboration only.
Where we fit
We sit beside the platform. We do not replace it.
Configuration vs production drift
Ema's strength is fast configuration: an enterprise can stand up a new AI employee conversationally, with the right governance and policy logic baked in. AgentStatus answers the next-layer question: a month after that AI employee is deployed across 200K users, is it still answering the way it was configured to, across regions, policy variants, and model updates?
Inside-out compliance vs user-side evidence
ISO 42001 and SOC 2 give an enterprise a strong baseline: the platform is governed correctly. Distributed validate traffic provides a different layer of evidence: that the deployed agent's actual outputs continue to match the gold standard, day over day. Two layers of trust, not one.
Global execution footprint
2,500+ nodes across 70 countries is the proof we are not 'synthetic from a single cloud region.' For enterprise AI employees serving hundreds of thousands of users across geographies, HR, legal, customer support, finance, it matters that residential observers validate inbound reach from where employees and customers connect. Multi-geo is access and last-mile latency, not answer localization by probe IP.
Partner-friendly integration posture
We do not assume we can 'discover' Ema customers the way some web-widget vendors can be scraped. Credential-based surfaces (agent endpoints, sandbox AI employees, customer-approved monitoring) are the right model, aligned with the on-prem and air-gapped deployment options Ema supports for its most security-sensitive customers.
The split
How the work divides
Your platform
- • Generative Workflow Engine
- • EmaFusion 2T+ model
- • Pre-built AI employee library
- • ISO 42001 / SOC 2 / HIPAA / GDPR
- • On-prem & air-gapped deployments
Outcome
System of record
Dashboards, exports, lifecycle tools, and orchestration remain yours. We do not replace that surface.
AgentStatus
- • Continuous validate traffic
- • Expected-answer checks & drift detection
- • Multi-turn / multi-agent journeys
- • Real-network execution evidence
- • 2,500+ nodes across 70 countries
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.
Proof of scale
Auditable scale metrics
In about two months, we have executed on the order of 18 million validate runs across the network. We also maintain on the order of 6,000 agent records in our system, meaning rows/configurations we track, including evaluation and pipeline agents, not "6,000 paying customers."
If helpful, we can share stricter production-only definitions under NDA.
What we are not claiming
We are an independent layer that runs alongside your stack.
We are not a replacement for Ema's Generative Workflow Engine, EmaFusion model, or compliance posture. We are an independent layer that can coexist with them, and, where useful, help enterprises correlate user-side validate outcomes with inside-out workflow execution, so leaders deploying AI employees at 200K+ scale have continuous evidence the deployed agents are still behaving the way they were configured to.
What we'd like from this conversation
These three asks would move a pilot forward.
A 2-week sandbox pilot
A sandbox AI employee (claim validation, compliance analyst, agent QA, or prior authorization), a set of agreed scenarios with expected answers, and a 2-week evaluation window. No production traffic, no enterprise data. At the end you get a written report of what we tested, what passed, and what drifted.
Security and procurement posture
How AgentStatus should connect in a way that satisfies enterprise security reviews under ISO 42001, SOC 2, and HIPAA. Data handling, least privilege, audit evidence, and clear test-traffic boundaries (including for on-prem and air-gapped deployments).
Where independent proof is most useful
Whether the right starting point is Ema-internal QA, a joint enterprise scenario in financial services or healthcare, or both.
Ema helps enterprises build and operate AI employees across every role in the organization.
AgentStatus helps those same enterprises prove, continuously, that those AI employees behave the way policy and customers require, globally, with evidence that holds up under scrutiny.
Contact·dulra@carmel.so·roman@carmel.so
Metrics are stated with explicit definitions: validate runs are scheduled executions over ~two months; agent records are database rows, not revenue customers. Public Ema references above reflect Ema's public product pages, compliance documentation, and Series A funding announcement as of the date of this note.