Skip to content

AI Managed Checks FAQ

Answers to common questions about how AI Managed checks behave, how the conversion to Authored works, and what to expect from future Profile runs.

What is an AI Managed check?

A quality check that Qualytics AI creates for you based on the statistics gathered during a Profile operation. Qualytics AI runs machine-learning models against your field-level metadata and generates rules that fit the data, so you do not have to write them by hand.

How do I get AI Managed checks?

Run a Profile operation on the datastore. Once the operation completes, the AI Managed checks it produced appear in the Checks list. Click any check to open its overview — Owner shows AI Managed while Qualytics manages the rule, and the History panel records the original creation by Qualytics.

What does the AI badge mean?

The purple AI badge (four-point star icon) marks a check that is currently managed by Qualytics AI. Future Profile runs may refresh or remove it. Hovering the badge shows the tooltip "Authored by AI and is continuously tuned to your observed data." As soon as you change a property that affects how the check evaluates the data, the badge is replaced with Authored and the check is no longer touched by Profile.

Editing and Conversion

What edits turn an AI Managed check into an Authored one?

Changing the rule properties, coverage, filter clause, or fields the check applies to. Any of those four edits triggers the conversion as soon as you save.

What edits keep the check AI Managed?

Editing the description, tags, additional metadata, anomaly message field, owner, anomaly assignee, or the check status (Active/Draft). These edits are organizational and do not change how the check evaluates the data.

Will the platform warn me before converting?

No. The conversion happens silently as soon as you save. The visible signal is the AI badge being replaced with the Authored badge.

I edited a property by mistake. Can I undo the conversion?

There is no one-click "revert to AI Managed" action. If you want the check back in the automatically managed set, delete it and run a new Profile operation — Qualytics AI will recreate an AI Managed version of the rule if the data still supports it.

Does converting an AI Managed check change who owns it?

If the AI Managed check was originally owned by Qualytics, ownership is transferred to you, the user who edited it. The check then shows up in lists like "checks I own".

Profile Behavior

What does the next Profile run do to my AI Managed checks?

For checks that are still AI Managed, Profile does one of three things:

  1. Keeps the check as is, if the rule still fits the data.
  2. Adjusts the check, if Qualytics AI refines its understanding of the rule.
  3. Removes the check, if Qualytics AI no longer detects the rule with enough confidence.

Profile never modifies or removes an Authored check.

Will Profile re-create an AI Managed check I deleted?

Yes, if the data still supports the rule. Deleting an AI Managed check is a clean reset — the next Profile run can re-generate it.

How often should I re-run Profile?

Profile runs are typically scheduled — daily, weekly, or after significant data loads — so AI Managed checks stay aligned with how the data actually behaves. See the Profile Operation documentation for scheduling guidance.

API

Is there a separate API for AI Managed checks?

No. AI Managed and Authored checks share the same Quality Checks API. Each check returned by the API has an inferred boolean that tells you which it is (inferred: true = AI Managed; inferred: false = Authored). The field name inferred is preserved for backward compatibility. See the API page for examples.

Does editing an AI Managed check via the API also convert it?

Yes. The conversion rules are the same in the UI and in the API. Sending properties, coverage, filter, or fields in a PUT request flips inferred to false on save.

Permissions

Who can view AI Managed checks?

Anyone with team access to the datastore — including Reporter and Viewer team permissions.

Who can edit AI Managed checks?

The required permission depends on the field. Drafter is enough for description, tags, additional metadata, anomaly message field, owner, anomaly assignee, and status. Editing rule properties, coverage, filter, or fields requires Author or above.

For the full breakdown, see Permissions.