The autonomous CMDB. Discovers, reconciles, and corrects itself.

An autonomous CMDB is a Configuration Management Database that runs itself. It discovers assets on its own, reconciles them into one record per real-world thing, corrects drift without a human in the loop, and surfaces changes the moment they happen. Where a traditional CMDB asks the team to be the accuracy mechanism, an autonomous CMDB makes accuracy a property of the system itself.

A live CMDB is a Configuration Management Database that stays accurate without manual upkeep. It is the visible result of an autonomous operating model: continuous discovery, automatic reconciliation, and a configuration record that reflects what is on the network today rather than a snapshot from the last audit.

They describe the same thing from two angles. "Live" is the property the user sees - the CMDB is current, all the time. "Autonomous" is the operating model that produces it - the CMDB does the work of staying current on its own. Every autonomous CMDB is live; an asset register that updates itself is by definition both.

A living CMDB has a heartbeat. Kovira's discovery surfaces (ATLAS for endpoints, VECTOR for the network) observe the environment continuously, reconcile what they see against the existing record, and update the CMDB in step with the environment. The same way a living thing has vital signs, a living CMDB has a measurable, predictable pulse.

Two layers of the same nervous system. The heartbeat is the steady cadence of base discovery - included in every plan and good enough that the configuration record is always trustworthy. The reflex is the optional LIVE add-ons (ATLAS LIVE and VECTOR LIVE) that push changes into the CMDB the instant they happen, for teams whose automations need to react in seconds. Both layers feed the same single record per asset.

Fast enough that the configuration record is always trustworthy. On the heartbeat tier the CMDB is continuously refreshed by discovery; on the reflex tier (LIVE add-ons) changes are reflected near-instantly. The exact cadence is tuned automatically to the size and pace of your environment - we don't expose specific intervals because the goal is for you to never have to think about them.

When the time between a real-world change and the CMDB knowing about it has to be near-zero. Typical fits: security workflows that fire on posture drift; MSPs who want a tenant's network change to surface in their console immediately; teams running automation chains that depend on second-fresh inventory. For most IT teams, the heartbeat is enough; LIVE is the upgrade for teams whose workflows assume reflex-speed signal.

Autonomy and AI are related but not synonymous. AI features (correlation, anomaly detection, dependency inference) accelerate the autonomous loop, but the autonomous part is the operating model: the CMDB observes, decides, and corrects without waiting on a human approval queue. Kovira combines both - automated discovery as the base, smart reconciliation on top, and an explicit audit trail so the autonomy is observable rather than mysterious.

A traditional CMDB is a manual system of record - accurate when someone keeps it accurate, stale the moment the environment moves faster than the people maintaining it. An autonomous CMDB closes that gap structurally. The system is responsible for accuracy, not the team, so the configuration record is current by default rather than current when someone gets around to it.

CMDB drift is structural. Environments change faster than people can document them: devices appear during change windows, software rolls out across the fleet, laptops get swapped, and someone is always meant to record it. When the update path is manual, it loses to whatever is on fire. Within a few months the CMDB no longer matches the network, and within a year nobody trusts it.

CMDB drift is the gap between the configuration record and reality. Autonomy prevents it by removing the manual step. The CMDB is rebuilt against reality continuously - new assets are detected, decommissioned ones get a last-seen timestamp, attribute changes flow in automatically. There is no audit cycle to fall behind on because there is no audit cycle at all.

Kovira observes the environment through built-in discovery (ATLAS endpoints, VECTOR network, Microsoft 365 directory and Intune), reconciles every signal into one record per asset, and self-heals duplicates without human intervention. Workflows fire against the current state, change reviews see real dependencies, and the configuration record is the source of truth rather than a record to be reconciled with one.

For accuracy, no - the autonomy handles that. For decisions, yes. Approvals on impactful changes, the policy of which signals to trust most, and the boundaries of automated workflows still belong to people. Autonomy removes the toil; it doesn't replace judgement. Kovira surfaces every autonomous decision in an audit trail so the human-in-the-loop knows what changed without having to dig.

A living, autonomous CMDB is the data foundation underneath ITSM, not a replacement for it. Kovira is both - an autonomous CMDB and an ITILv4-aligned ITSM platform - so incident management, change management, and impact analysis all run against configuration data that is current rather than against records nobody trusts.

Small teams benefit the most. The smaller the team, the harder it is to dedicate someone to manual CMDB upkeep, and the more painful drift becomes. Kovira's free tier includes the full autonomous CMDB so a single technician can maintain a current configuration record without the operational overhead.

Change approvals, impact analysis, incident triage, and compliance reporting all assume the CMDB is right. When it is not, decisions are made on yesterday's picture of the environment. An autonomous CMDB removes that risk - reviewers see current dependencies, auditors see a record that matches reality, and incident responders trust what they are looking at.

Yes. The autonomous, living CMDB - including ATLAS and VECTOR on the base heartbeat tier - is included in every Kovira plan, including the free tier. The LIVE add-ons (ATLAS LIVE and VECTOR LIVE) are optional reflex upgrades for teams who want push-on-change updates instead of continuous scheduled discovery.

Endpoints, servers, network gear (with topology and L2 link enrichment), users, groups, software, contracts, vendors, locations, racks, domains, APIs, passwords, and Microsoft 365 surfaces (Entra ID identity, Intune devices and policies). Every CI lives in one consistent schema with a full version history, so autonomous updates are auditable and reversible per change rather than per-record snapshots.