Abnormal AI’s 2025 updates expand human threat coverage

At a glance

What this is: Abnormal AI's 2025 update bundle focuses on behavioural detection, risky Microsoft 365 posture, and expanded coverage for email and collaboration threats.

Why it matters: For IAM and security teams, the signal is that user-facing identity risk now spans configuration drift, collaboration channels, and behavioural abuse, not just inbound email.

By the numbers:

• Abnormal says its core models use 50 percent more features, enabling proactive identification of tens of thousands of additional attack campaigns each week without increasing false positives.
• Abnormal says its expanded phishing coverage reduced missed detections by 30 percent.

👉 Read Abnormal AI's 2025 overview of human threat detection and remediation updates

Context

Human-targeted threats are no longer confined to the inbox. Attackers now use calendar invites, collaboration tools, and behavioural lures to get around legacy filters that were built for more predictable message patterns.

That shift matters to IAM and security programmes because email security, collaboration security, and identity posture are converging. Once risky Microsoft 365 settings, excessive permissions, or misdirected messages become part of the attack path, the problem is as much governance as it is detection.

Key questions

Q: How should security teams defend users across email, calendar, and chat channels?

A: They should treat those channels as one collaboration risk surface and align detection, alerting, and remediation across all of them. If email is protected but calendar invites or Teams messages are not, attackers will simply move to the least monitored path. Consistent policy and shared response ownership matter more than isolated tools.

Q: Why do behavioural models matter for phishing defence?

A: Behavioural models help security teams catch attacks that do not match known signatures, especially personalised phishing and account abuse. They look for shifts in sender behaviour, message patterns, and context, which makes them better suited to adaptive campaigns than static rule sets alone. That reduces missed detections without relying only on keyword filters.

Q: What breaks when Microsoft 365 permissions and settings are left unmanaged?

A: Attackers inherit a much larger blast radius. Excessive permissions and risky settings make it easier for a phishing or collaboration lure to become account abuse, data exposure, or lateral movement. When posture management is missing, the environment itself becomes part of the attacker’s pathway.

Q: Who is accountable when collaboration-channel attacks lead to data exposure?

A: Accountability usually sits across security operations, identity governance, and platform administration because the failure spans detection, permissioning, and user protection. If Teams, calendar, and email are managed separately, no single team sees the full path. Mature programmes define ownership for both channel protection and tenant posture.

Technical breakdown

Behavioural detection in email security

Behavioural detection models look at patterns in sender behaviour, message timing, content signals, and historical context rather than relying only on signatures or static rules. That matters because modern phishing is increasingly personalised and fast-moving, so static filtering leaves gaps when the lure has never been seen before. Abnormal says its 2025 model changes expanded the feature set and multilingual understanding, which is the kind of tuning that aims to catch broader campaign patterns without increasing alert noise. The technical point is not just better classification. It is using context to separate ordinary communication from suspicious behavioural shifts.

Practical implication: security teams should measure whether their controls can detect new campaigns without creating false-positive fatigue.

Microsoft 365 posture and permission drift

Security posture management in Microsoft 365 focuses on the configuration layer around users, not just the messages they receive. Risky settings and excessive permissions create ready-made breach points because an attacker does not need to defeat every control if the environment is already permissive. This makes configuration drift an identity issue, not only an admin concern, because the effective security boundary is shaped by who can change settings, who can approve exceptions, and which permissions persist. In practice, the attack surface includes the governance state of the tenant as much as the content arriving in it.

Practical implication: teams should treat M365 permission review and configuration drift as part of identity governance, not a separate hygiene task.

Collaboration-channel abuse in Teams and calendar invites

Attackers move into collaboration channels because users trust those surfaces and defenders often monitor them less aggressively than email. Calendar invites can carry malicious content or timing-based lures, while Teams messages can be used for social engineering or payload delivery in the flow of daily work. Once a platform extends protection into these channels, the challenge becomes managing response across multiple communication systems with consistent policy and remediation logic. The technical issue is channel expansion: the attacker path follows the user, not the inbox.

Practical implication: security operations should map remediation and alerting across email, calendar, and chat so one channel does not become an unmonitored bypass.

Threat narrative

Attacker objective: The attacker aims to turn trusted human communication channels into a low-friction path for compromise, data exposure, or further account abuse.

1. Entry occurs when attackers weaponize trusted communication channels such as calendar invites or Teams messages to bypass legacy email filters and reach users.
2. Escalation occurs when behavioural context is absent or weak, allowing socially engineered content, risky links, or misdirection to blend into normal collaboration activity.
3. Impact occurs when the lure triggers credential theft, malicious interaction, or accidental data exposure through misdirected email or compromised collaboration workflows.

Breaches seen in the wild

• ASP.NET machine keys RCE attack — 3,000+ exposed ASP.NET machine keys enabled remote code execution.
• DeepSeek breach — DeepSeek breach exposed 1M+ log lines and sensitive secret keys.

Read our 52 NHI Breaches Analysis report for a comprehensive view of breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Human threat protection is now a multi-surface governance problem, not an inbox problem. The article shows that email, calendar, Teams, and Microsoft 365 configuration are now part of the same control plane. That means the organisation is no longer defending a single channel, but a mesh of user touchpoints where trust can be manipulated. Practitioners should expect governance gaps to appear wherever detection and response do not follow the user across surfaces.

Behavioural context has become the differentiator between noise and actionable human risk. A model that can reduce missed detections while preserving false-positive rates is not just about better machine learning. It is about preserving analyst trust in the signal while attacks become more personalised and harder to classify with static rules. Security leaders should evaluate whether their current stack can still separate normal collaboration from manipulated intent.

Configuration drift and identity context now belong in the same conversation. Excessive permissions and risky Microsoft 365 settings are no longer background hygiene issues. They shape how far a successful lure can travel once a user is engaged, which means posture management and identity governance need shared ownership. The implication is that the effective blast radius is defined by permissions as much as by detection speed.

Channel expansion is the new normal for human-targeted attacks. Attackers follow the place where work happens, so controls that stop at email will keep missing the next trusted path. Calendar and Teams protection show where the market is heading: security programmes must govern communication surfaces as a unified exposure layer, not as isolated tools. Practitioners should re-baseline control coverage across the collaboration stack.

Behavioural trust debt: The article illustrates a growing gap between the speed of human-targeted attacks and the assumptions embedded in legacy filters. Those controls were designed for stable patterns, not for adaptive campaigns that change payload, channel, and timing to match user behaviour. Practitioners should treat that mismatch as a structural governance issue, not just a detection tuning problem.

From our research:

• Two-thirds of enterprises have endured a successful cyberattack resulting from compromised non-human identities, with a quarter encountering multiple attacks, according to The 2024 ESG Report: Managing Non-Human Identities.
• From our research: 72% of organisations have experienced or suspect they have experienced a breach of non-human identities, including 46% confirmed and 26% suspected, according to the 2024 ESG Report: Managing Non-Human Identities.
• Forward pivot: For a deeper lifecycle lens on access, rotation, and offboarding, see the NHI Lifecycle Management Guide and align communication-channel governance with identity controls.

What this signals

Behavioural detection is moving from point capability to governance expectation. As attackers shift across email, calendar, and collaboration tools, teams need visibility that follows the user rather than the inbox. The practical signal is that detection accuracy now has to be judged alongside response reach and ownership across platforms.

Configuration drift is now part of the identity attack surface. When risky Microsoft 365 settings and excessive permissions remain visible but unmanaged, attackers do not need a new exploit path. They need only one human interaction to turn permissive governance into exposure.

More than 1 in 5 non-human identities are insufficiently secured, according to our 2024 ESG Report: Managing Non-Human Identities. That scale matters here because collaboration and tenant controls increasingly determine whether human-targeted attacks stay contained or turn into account abuse.

For practitioners

• Extend detection beyond the inbox Map email, calendar, and Teams into one response workflow so malicious content can be identified and removed across channels before users continue interacting with it.
• Review Microsoft 365 permission drift Continuously surface risky settings and excessive permissions in Microsoft 365, then assign ownership for remediation so configuration issues do not sit outside identity governance.
• Measure missed-detection reduction against false positives Track whether new behavioural models improve campaign catch rates without increasing analyst workload, because precision matters as much as coverage in human threat defence.
• Add misdirected-email guardrails Use behavioural context to flag likely recipient mistakes before send, especially where sensitive data or external recipients are involved.

Key takeaways

• Human-targeted attacks now span email, calendar, Teams, and Microsoft 365 posture, so channel-specific defence is no longer enough.
• Behavioural context improves phishing detection by reducing missed threats without forcing a false-positive trade-off that overwhelms analysts.
• Identity governance and posture management must share responsibility for collaboration risk because permissions and configuration shape attacker reach.

Key terms

• Behavioural Detection: A detection approach that looks at how accounts, senders, or messages behave over time instead of relying only on static signatures. It is useful when attackers personalise lures or shift channels, because the abnormal pattern often appears in context rather than in the content alone.
• Security Posture Management: Continuous assessment of configuration, settings, and permissions that shape an environment’s effective exposure. In identity-heavy environments, posture management shows where excessive access or risky defaults can turn a routine interaction into a breach path.
• Collaboration Channel: Any work surface where people exchange messages, files, or invitations, including email, chat, and calendar systems. These channels are now part of the identity attack surface because trust in them is often higher than the trustworthiness of the content they carry.
• Misdirected Email: A message sent to the wrong recipient, often because a sender selects an incorrect contact, group, or thread. When the content is sensitive, the mistake becomes a data exposure event, so prevention depends on behavioural context as well as recipient validation.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity lifecycle are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are responsible for identity security strategy or NHI governance in your organisation, it is worth exploring.

This post draws on content published by Abnormal AI: Key insights from its 2025 platform updates for human threat protection. Read the original.