How Reqursor fixes unexpected changes with AI
The full flow from an unexpected change being spotted to an AI diagnosis, a proposal, your confirmation, and the fix going live.
Understanding unexpected changes covers how Reqursor spots changes nobody asked for: gaps between a store and its saved setup (its declaration). Spotting a gap is only half the story. This page covers the other half: how a spotted gap actually gets fixed.
The full path, start to finish
An unexpected change is spotted
A check, either on a schedule or when you ask for it, finds a mismatch between what's saved as the store's setup and what's actually running.
The AI reads the evidence
Not just "something's wrong": the AI looks at exactly what differs, whether it appeared right after a recent change, whether it's actually breaking anything customers can see, and whether other stores in the same group show the same issue.
The AI produces a proposal
A specific recommendation: what it would do, which parts of the store and which stores are affected, a plain-language explanation of what changed and why, how risky the fix is, and what else you could choose instead.
You review and confirm
Through the dashboard or in conversation, see Fixing unexpected changes from the dashboard or in chat. Nothing runs without this step.
Reqursor puts the fix live
The confirmed action goes through the same plan, apply and check process as any other change.
A follow-up check confirms it worked
The same kind of check that spotted the problem runs again, to verify the store now matches its saved setup.
What the AI actually looks at
| Source | What it tells the AI |
|---|---|
| The unexpected change itself | Exactly what differs: for example, that the homepage's primary color no longer matches the saved setup |
| Recent history | Whether the change appeared right after a specific recent change |
| Health signals | Whether it's actually breaking something customers can see, or is purely cosmetic |
| Your other stores | Whether other stores in the same group show the identical issue |
What a proposal actually tells you
A proposal isn't a vague alert. It reads like a specific recommendation: "Put Store X back to its saved setup to restore the primary color to navy. Someone changed it to red outside Reqursor." Alongside that, you see how risky the fix is (routine, moderate, or higher-risk) and what alternatives you could choose instead.
Nothing happens without your OK
Reqursor does not fix problems on its own. Every proposal, no matter how routine, goes through the same preview-and-confirm step as any other change.
Why this matters for your agency
- You get a recommendation, not a raw alert. The AI does the investigating before it ever reaches you.
- Risk is visible up front. You know whether a fix is routine or worth a closer look before you approve it.
- The same process works at any size. One store or a hundred, unexpected changes are looked into and fixed the same way.
Where to go next
Understanding unexpected changes
How unexpected changes are found in the first place, on a schedule and across all your stores.
Fixing unexpected changes from the dashboard or in chat
Where you actually see and confirm the proposal this process produces.
Six ways your store's setup gets written
Where AI fixes fit among the other ways a store's saved setup gets written.
Why changes made outside Reqursor are caught
Why editing a store directly, outside Reqursor, is always treated as an unexpected change, never as a new saved setup.
Spotting unexpected changes that keep coming back
How Reqursor recognizes when the same unexpected change keeps happening, on one store or across a group, and asks you instead of repeating the same fix blindly.