TL;DR
Learn why Cayosoft has been named a Sample Vendor for AI agent action rollback in Gartner’s Hype Cycle for AI Governance Technologies, 2026.
We’re excited to announce that Gartner has named Cayosoft a Sample Vendor in two categories in the Hype Cycle for Backup and Data Protection Technologies, 2026: Identity Recovery and AI Agent Action Rollback. We’re proud to be recognized in both categories, but what makes this especially meaningful is the connection between them. We believe that together, they reflect a common goal: helping organizations recover quickly, confidently undo unintended changes, and maintain control in increasingly complex identity environments.
When identity goes down, nobody can log in, applications stop working, tickets spike, and every other recovery effort waits for identity to come back first. That is true whether the cause was ransomware, an administrator’s mistake, or an autonomous agent doing what it was told with the permissions it was given. Agent-driven incidents may arrive faster and at greater scale, but the directory sees the same event: something changed, and someone has to put the right part of it back.
Cayosoft Guardian addresses both with the same change history and the same rollback capabilities, regardless of whether the change was made by a human administrator, a service account, or an AI agent.
The hard part of identity recovery is not having a backup. It is knowing what to put back. Every recovery decision comes down to two questions: how much needs to be reversed, and how far back should you go?
Identity recovery has been treated as the backup team's problem
Gartner defines the category as:
“Identity recovery solutions enable secure backup and restoration of an organization’s identity and access management (IAM) data and configurations across platforms such as Microsoft Active Directory, Microsoft Entra ID and Okta Identity Cloud. They support rapid, granular recovery from cyberattacks, human errors and misconfigurations by restoring critical IAM components such as identities, roles, permissions, policies and authentication settings to a clean, trusted state, minimizing disruption.”
Most identity programs have spent the last decade on prevention: Conditional Access, privileged access management, and MFA everywhere. Recovery was often left to the backup team, and the backup team’s tools were built to recover data, not identities.
Gartner expects that mindset to change:
“By 2030, 70% of organizations will prioritize the recoverability of identity systems alongside preventive identity and access management (IAM) controls, up from less than 15% in 2026.”
We believe this shift reflects a growing recognition that prevention alone is not enough. Organizations also need a way to understand what changed, determine what should be reversed, and restore identity systems quickly when prevention fails.
A point-in-time restore does not match most identity incidents
Identity incidents tend to fall into one of two categories: a single bad change to a single object, or a catastrophic event that compromises an entire forest. Recovery tools designed for only one of those scenarios are poorly suited for the other. In our opinion that distinction is reflected in Gartner’s recommendations for the category:
“Assess vendors for automated full and granular recovery capabilities without requiring full environment rollback.”
The need for precision becomes even more important in the age of AI agents. As Gartner notes:
“Standard backup and restore are too blunt for many agent-driven incidents. They recover to a point in time, but they do not isolate which changes were caused by a specific agent or preserve unrelated states with targeted precision.”
For smaller incidents, native Microsoft capabilities provide only partial coverage. Administrators may be able to use Active Directory Recycle Bin where it is enabled, GPO backups through Group Policy Management Console, Entra ID’s 30-day soft delete for a limited set of object types, or Microsoft Entra Backup and Recovery for daily property-level restores. But coverage remains inconsistent. On-premises synchronized objects are excluded from recovery, retention is limited to seven daily backups, and several object types cannot be restored at all. Most importantly, these tools can restore data, but they do not identify which specific change should be reversed.
At the other end of the spectrum, a full forest recovery is a major operational undertaking. Microsoft’s Active Directory Forest Recovery Guide outlines a lengthy process: restore one writable domain controller per domain in an isolated network, beginning with the forest root; reset privileged credentials and the krbtgt account following an intrusion; seize operations master roles; and clean up metadata for every domain controller that is not recovered. Throughout that process, no writable domain controller is available to issue new authentications, meaning password changes fail and any service requiring a fresh Kerberos ticket or LDAP bind is disrupted.
That level of recovery effort may be necessary after a catastrophic compromise. It is not an appropriate response to a single incorrect group membership, misplaced license assignment, or unintended attribute change.
The reality is that most identity incidents require precision, not a point-in-time rollback. Organizations need the ability to identify what changed, determine the scope of impact, and reverse only the affected objects, attributes, policies, or permissions without disrupting everything else.
Choosing what to put back requires knowing what changed
The second question is how far back to go, and in a compromise, the most recent backup is often the wrong answer. Attacker persistence, unauthorized privilege grants, tampered policies, and backdoor accounts are all directory state, so the backup holds them alongside everything legitimate. From Gartner’s user recommendations for the category:
“Validate recovery points before restoring roles, permissions and policies to prevent privilege escalation or reinfection.”
Validating a recovery point means knowing what changed and when. A common example: a service account with delegated permissions broader than intended is used to add an account to a privileged group, and the change surfaces days later during an unrelated investigation. Which of the several hundred changes since then belong to the intrusion, and which are legitimate work? Without a change history that predates the incident, the compromise window is a guess, and selecting a restore point becomes a choice between going back far enough to feel safe and losing legitimate changes along with it.
Agents make that history even more important. An AI agent does not need a privileged directory role to create a recovery problem, and under Microsoft Entra Agent ID, it will not have one, since tenant-wide write permissions and the high-privilege directory roles are blocked for agent identities. What agents do get is group memberships, access packages, app role assignments, and the delegated permissions of the user they act for. An agent built as a conventional app registration gets whatever permissions an administrator consented to.
Changes made with those rights land on the same objects an administrator edits through Microsoft Graph, often at a rate no one reviews change by change. The question is the same when the actor changes: which changes were this agents, and what has to go back?
A chosen recovery point still must be executable
Deciding what to restore is only useful if the restore can actually run. Attackers go after the recovery path, and recovery assets that share production identity, credentials, or management planes inherit the same blast radius. The recovery console signs in against the directory that is down. The backup repository is reachable with administrator credentials the attacker already holds, and the recovery documentation lives in a system that depends on what is being recovered. Isolation is what leaves the decision executable.
Rehearsal is the other half of the equation. A recovery path that has never been tested is still a hypothesis.
From Gartner’s user recommendations for the category:
“Prioritize solutions that enable safe, automated testing of IAM recovery using isolated or nondisruptive validation capabilities.”
How Cayosoft Guardian supports that decision
- Change history that identifies the compromise window.
Guardian monitors Active Directory, Entra ID, Microsoft 365, and Intune continuously and keeps before-and-after change history with privileged activity context, tracking service accounts and other non-human identities alongside human administrators. That history is what makes a recovery point a choice rather than a guess. - Recovery at the granularity of the incident.
Object- and attribute-level rollback across Active Directory and Entra ID handles the single bad change: one group membership, one attribute, one deleted object, without reverting everything else that happened that day. It also recovers objects Entra can no longer restore itself, repairing the entire relationship graph. - Full recovery as a cutover, not a rebuild.
Guardian pre-builds and continuously maintains a clean standby Active Directory forest, with forest, domain, DNS, SYSVOL, and FSMO recovery readiness prepared and refreshed at every backup instead of executed during the outage. Identity services come back in minutes rather than the hours or days a rebuild takes. Paradigm Technica’s independent technical and economic validation found Cayosoft at least 99% faster than the closest AD recovery alternative. - A recovery path outside the blast radius, validated before it is needed.
Standby forests can live in Azure, in AWS, or offline, on dedicated networks, with immutable WORM storage and cloud-only administrators that share no credentials with production, and the cloud-side recovery environment can use a separate Entra ID tenant. Because the cloud-hosted standby runs independently of production domain controllers, DNS, and system state backups, it survives an incident that takes the primary Cayosoft Guardian server with it. Guardian continuously maintains and validates the standby and the recovery process, so the team knows recovery works before it needs it.
For any team planning identity recovery this year, the question worth asking is not whether a backup exists. It is whether you can tell what changed, reverse exactly that much, and can confidently restore to a trusted state, and get the same answer whether the change came from an administrator, a script, or an AI agent.
The full Gartner Hype Cycle for Backup and Data Protection Technologies, 2026 is available through Gartner subscription access. [Read full report]
Gartner, Hype Cycle for Backup and Data Protection Technologies, 2026, Michael Hoeck, 28 July 2026.
GARTNER and HYPE CYCLE is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved. Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner’s research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.
See Cayosoft in Action
Cayosoft is recognized by Gartner as an ITDR solution provider and provides solutions that make identities more resilient to attacks and guarantee a fast forest recovery, if needed. Learn how Cayosoft Guardian facilitates granular change tracking, post-breach analysis, and long-term AD security improvements. Schedule a demo to see the capabilities in depth.