Most have never been reviewed.
AIRM continuously monitors every AI agent and non-human identity in your Microsoft 365 tenant. It scores risk, maps blast radius, detects rogue behaviour, and produces compliance evidence for EU AI Act, DORA, ISO 42001 and other frameworks. Detection through to guided response in one platform.
AI agents and non-human identities now outnumber human users 100 to 1 in most Microsoft 365 environments. Almost no organisation has visibility into them.
Copilot connectors, Power Automate flows, and third-party AI tools operate continuously with permissions nobody reviews. If compromised, the damage is catastrophic.
Service principals accumulate over time with aged credentials, no owners, and permissions far beyond what they need. A quiet app with write access to your entire tenant.
EU AI Act, DORA, ISO 42001: regulators are asking about AI agents and non-human identities. Most organisations can't answer those questions yet.
AI adoption is accelerating faster than security teams can govern it. 92% of security leaders are concerned about AI agents and their impact on security. AIRM was built to answer these questions, and to act on the answers.
Your board and CEO are pushing AI adoption. You've been asked to sign off on it. But 69% of enterprises already have AI agents running in production. Most were deployed without security review, with permissions nobody explicitly granted.
AIRM gives you a full inventory of every AI agent and non-human identity in your Microsoft 365 environment. You can answer "what AI do we have and what can it access?" before you answer "are we ready?"
3 in 4 CISOs have discovered unsanctioned AI tools already running in their environments. Many were granted OAuth tokens and elevated permissions plugged directly into enterprise systems, outside any standard provisioning workflow.
AIRM detects every service principal and AI agent in your tenant, whether sanctioned or not. Unreviewed identities are flagged immediately. You see their permissions, blast radius, and compliance implications together.
An AI agent your organisation legitimately approved can still be compromised. Its code can be updated with malicious logic. It can be silently granted new permissions. Traditional security tools watch for human behaviour anomalies. They weren't built to catch this.
AIRM's Blast Radius Analysis maps exactly what an attacker would reach if any identity were compromised. The map is built from actual granted permissions, not assumed activity. You see the worst case before it happens.
Regulators are now asking specific questions about AI governance. The EU AI Act's high-risk obligations phase in through 2026. DORA has been in force since January 2025. Most organisations cannot yet produce evidence of AI agent oversight.
AIRM automatically maps every finding to 11 compliance frameworks including EU AI Act, ISO 42001, and DORA. It generates per-framework compliance reports you can hand to auditors and regulators.
CISOs are under pressure to enable AI innovation. Saying yes without visibility is the fastest way to lose control. Organisations that establish agent inventories and privilege policies now will be the ones in control when something goes wrong.
AIRM gives you the visibility layer that makes "yes" safe. Monitor every AI agent continuously, detect permission drift, get alerted when behaviour changes, and demonstrate governance to the board. One platform.
You don't have to choose between enabling AI and securing it. AIRM gives security teams the visibility, intelligence, and evidence they need to say yes to AI with confidence.
Start free trial — 14 daysAIRM continuously scans your Microsoft 365 tenant. It discovers every AI agent, service principal, and non-human identity, classifies them, and tracks every change across every scan.
AIRM scores every identity on two independent dimensions: what it's doing right now, and how bad it would be if it were compromised. A quiet app with dangerous permissions is your biggest threat.
The Blast Radius Map visualises the potential impact of a compromised identity based on its granted permissions, not historical activity. Every Service Principal with permissions shows a populated map from day one.
Directory.ReadWrite.AllCRITICALApplication.ReadWrite.AllCRITICALRoleManagement.ReadWrite.AllCRITICALThe Identity Graph maps the full web of trust relationships across your Microsoft 365 tenant. It shows AI Agents, Non-Human Identities, humans, and Microsoft first-party apps, and exactly who has permission to act on what. Three modes reveal different threat dimensions: the full Graph view, Attack Paths for lateral movement chains, and Temporal Analysis showing how the graph has shifted over time.
Compromise Simulation runs a permission-aware attack model against any AI Agent or Non-Human Identity. It translates raw Graph API scopes into human-readable attacker actions, scores the reachable blast, maps every accessible data type, and surfaces scoped remediation recommendations. No live data is touched.
If Copilot for Finance were compromised, an attacker would gain access to all mailboxes, email content, and 15 more data types. 13 distinct attacker actions available across 1 resource.
Mail.ReadWrite to specific mailboxes using resource-specific consent. Scope Files.ReadWrite.All using Sites.Selected where possible.AIRM maps every finding to the compliance frameworks your clients are accountable to, including the new AI-specific regulations coming into force now. Generate a per-framework compliance report in one click.
Just because an AI agent or service principal was sanctioned doesn't mean it stays safe. AIRM watches every identity continuously, and flags approved apps when they start behaving in ways they didn't before.
Unlike tools that take a point-in-time snapshot, AIRM's proprietary scoring engine accumulates behavioural intelligence with every scan. It builds a unique fingerprint for every identity in your environment. The engine has to know what normal looks like before it can detect what isn't. The longer AIRM monitors a tenant, the more precise its anomaly detection becomes. No static rule set or one-time assessment replicates this.
An AI agent your organisation legitimately approved and uses every day can be compromised. It can be updated with malicious behaviour. It can start accessing resources it never touched before. AIRM detects the delta, not just the initial state.
AIRM maintains awareness of service principals associated with known threat actors, compromised vendors, and malicious OAuth apps. When a known rogue Service Principal appears in your tenant, you're alerted immediately, not after the damage is done.
AIRM accumulates intelligence with every scan. The longer AIRM monitors a tenant, the more accurate its anomaly detection becomes. It knows exactly what normal looks like for every identity in your environment.
Organisations adopting Microsoft Copilot, custom Power Automate agents, and third-party AI tools need visibility into what those agents are doing, not just that they were approved. AIRM monitors sanctioned AI continuously so you know if it goes off-script.
Microsoft 365 E7 includes Agent 365, a significant step forward for AI governance. But it's built for Microsoft-native agents, at $99/user/month. AIRM covers the full Non-Human Identity attack surface, for every M365 tenant, at a fraction of the cost.
| Capability | MMicrosoft E7$99/user/mo |
AAIRMfrom $99/tenant/mo |
|---|---|---|
| AI Agent Visibility | ||
| Microsoft-native agent inventory (Copilot Studio, Agent Builder) | ✓ | ✓ |
| Third-party AI agent discovery (non-Microsoft apps) | — | ✓ |
| Shadow AI agent detection | Partial Entra Internet Access req. | ✓ |
| AI agent risk scoring (behaviour + blast radius) | — | ✓ |
| Non-Human Identity Security | ||
| Service principal inventory | Partial Entra admin only | ✓ |
| Non-Human Identity risk scoring and band classification | — | ✓ |
| Blast radius analysis (permission-based) | — | ✓ |
| Credential age monitoring and alerts | — | ✓ |
| Unowned identity detection and owner assignment | — | ✓ |
| Known rogue Service Principal detection | — | ✓ |
| Anomaly Detection | ||
| Approved agent behaviour monitoring | Partial Defender signals only | ✓ |
| Scan-over-scan anomaly accumulation | — | ✓ |
| Permission scope drift detection | — | ✓ |
| Behavioural baseline per identity | — | ✓ |
| Compliance & Reporting | ||
| AI governance framework mapping (EU AI Act, ISO 42001, NIST AI RMF) | — | ✓ |
| Per-framework compliance PDF reports | — | ✓ |
| MSP branded client reports | — | ✓ |
| Executive summary for non-technical stakeholders | — | ✓ |
| MSP & Multi-Tenant | ||
| Multi-tenant MSP dashboard | Partial Lighthouse — limited | ✓ |
| PSA integration (ConnectWise, HaloPSA, Autotask) | — | ✓ |
| Per-tenant pricing (not per user) | — | ✓ |
AIRM is built channel-first. MSPs get a multi-tenant dashboard, PSA integrations, and branded client reports. That's everything required to deliver Non-Human Identity security as a managed service.
See all client tenants at a glance. Risk scores, critical findings, and compliance signals across your entire client portfolio, in one view.
Native connectors for ConnectWise Manage, HaloPSA, and Autotask. Alerts become tickets automatically. No middleware required.
Every report carries your logo and company name. Executive summaries, Non-Human Identity risk reports, and compliance reports are all white-labelled at MSP Enterprise tier and above.
Plans run from $149/tenant/month (Basic) up to $1,599/tenant/month (Enterprise) — see pricing below. MSP partners receive volume discounts: 20% off at 5+ tenants (Silver), 30% off at 10+ (Gold), and 40% off at 25+ (Platinum), so you build healthy recurring margin into every client engagement.
AIRM doesn't just surface risk — it gets it into the hands of the people who can act. Smart alerting, PSA ticket creation, and outbound webhooks mean no finding sits unnoticed.
P1–P5 priority tiers automatically assigned based on risk severity. Deduplication prevents alert fatigue — the same finding doesn't fire twice. Auto-resolution when the underlying issue is fixed.
Snooze individual alerts for a defined period. Create suppression rules for known-good identities or accepted risks. Bulk acknowledge across multiple tenants from the MSP dashboard.
Alerts become PSA tickets automatically — no middleware, no Zapier, no API gymnastics. ConnectWise Manage, HaloPSA, and Autotask supported natively with structured ticket format and tenant context included.
Send AIRM findings to any system that accepts webhooks — SIEM, SOAR, Slack, Teams, custom dashboards. HMAC-SHA256 signed payloads for security. Configurable per event type and per tenant.
AIRM generates professional, branded PDF reports written in plain English — designed for both technical teams and non-technical stakeholders.
A consultant-quality summary of tenant risk posture — written in plain English for non-technical stakeholders. Includes risk narrative, top findings, and recommended actions.
Full operational summary of all AI agents — risk scores, anomaly flags, blast radius analysis, and scan results. Weekly reporting for security teams.
Complete risk profile of all non-human identities including behaviour risk, blast radius, attack paths, credential age, and per-identity scoring breakdown.
Per-framework compliance report with executive summary and per-control findings. Available for all 11 frameworks including EU AI Act and DORA.
Complete CSV export of all AI agents and Non-Human Identities for integration with your own reporting or ticketing workflows.
Every report carries your logo and company name. Hand clients a professional security report that looks like it came from your team, not a third-party tool.
AIRM maps every finding to 11 security and AI governance frameworks — including the new AI-specific regulations your clients are being asked about right now.
AI & Non-Human Identity Frameworks General Security Frameworks
Per-tenant pricing on annual commitment. All plans include a 14-day free trial — no credit card required.
Connect your first Microsoft 365 tenant in under 5 minutes. No agents. No complex setup. Just connect and go.
No credit card required · Agentless · Microsoft 365 only