What it corrects
Job change is where this started, and it is now one tool among several. Each kind of drift has one that corrects it. Every tool takes a record scope, keeps its own rules, and records what it wrote.
Find duplicate records and merge them.
The same company, three times. Duplicates detected and merged, with the surviving value chosen field by field.
Convert field values to canonical forms.
One city, five spellings. Cities, streets, websites, phone numbers and countries all put into one canonical form.
Fill missing fields from a data provider.
Fields nobody ever filled in. Industry, revenue and headcount fetched from a provider, to a ceiling you set.
Assign records to owners by strategy.
Records sitting on the wrong rep. Ownership routed by your territory policy instead of by hand.
Ask which contacts have moved, and queue each departure for review.
People who have already left. The departure caught, the contact archived, the person created at their new employer.
Find new people at a provider and add them to the CRM.
Companies you do not have yet. The gap between your CRM and your market, brought in and attached.
Apply a change across a filtered population.
A change that belongs to thousands of records. One instruction, one scope, one audit trail, instead of an afternoon in a spreadsheet.
Assign records to segments from rules.
Segments that live in someone's head. The rule written down once, applied to every record, and re-applied when the record changes.
Compute a score per record from weighted inputs.
A priority order nobody can explain. Weighted inputs, banded thresholds, and a score you can read back to the field that produced it.
Tell each record's owner something, over Slack.
A change nobody saw in time. The owner told about their own record, in the place they already work.
Check field values against an external source.
Values that were right once. A field checked against the world rather than trusted because it is filled in.
How they run
A tool is not a script. Each one is a saved, scoped operation you can read before it runs and read again after.
Every run names the records it covers, as a filter over your own fields. The same scope is what the preview counts and what the pass reads.
The decisions a tool makes are saved separately from the run that uses them, so the same policy runs twice without being retyped.
Every write is recorded with the value before it and the value after, attributed to the run that made it.
Where to start
The readiness scan runs the detection half of these tools against your own CRM and reports what it finds, before anything is changed.