Before using live customer data, confirm the required countries, entity types, source coverage, and recorded freshness in the source inventory. A product demonstration does not establish that your particular requirements are met. Assign a reviewer and someone who will handle incomplete checks.
01 · Prepare
Give each customer a stable reference.
Use the CSV columns customer_ref, name, type, date_of_birth, and country. References are optional for screening, but required when saving rows to your customer portfolio. Keep each reference unique within your organization and stable across later imports. Use individual or business; leave unknown optional identifiers empty rather than inventing them.
customer_ref,name,type,date_of_birth,country
sample-customer-001,Alex Sample,individual,1980-01-15,GBFictional input from the downloadable CSV. Alex Sample is not a customer or a real screening case.
Start with a small representative set before the full list. Resolve duplicate references and identity conflicts before saving. The batch accepts up to 1,000 rows. Inspect the upload contract and row-error sample.
02 · Screen
Inspect the row outcome before the match outcome.
| What you see | What to do next |
|---|---|
| Completed check; potential candidates | Compare the source evidence and identity fields. A similarity score is not a final decision. |
| Completed check; no candidates | Retain the result and its scope. It is a point-in-time check, not a compliance guarantee. |
| Invalid or failed row | Correct the input or handle the service failure. Treat the customer’s check as incomplete. |
| Partial batch | Inspect each row. Completed rows do not make the remaining failed rows clear. |
Keep a list of exceptions with an owner and next action. Successful saved live checks use the monthly allowance; a failed check is not a completed screening. Avoid repeating an entire completed batch just to correct a few failed rows.
03 · Review
Keep the reasoning with the candidate.
Compare names, aliases, birth dates, jurisdictions, and other available identifiers. Save a human disposition and a reason that another reviewer can follow. The review record retains the reviewer and timestamp alongside the evidence. Missing or conflicting identity fields need a reviewer’s judgment.
04 · Retain
Test whether someone else can retrieve the record.
Explicitly save selected completed live rows to your portfolio. Stable references link repeat imports to the same customer’s screening history. Review the per-row save outcomes and resolve conflicts; names alone are not a reliable identity key. Saving the same completed check again does not re-screen or charge again.
Ask a second reviewer to open the customer history, find the earlier check and its recorded reason, then open the screening PDF. Confirm the exported fields meet your team’s evidence needs. Historical competitor evidence is not automatically imported, and a previous decision does not automatically clear a new candidate.
Open the fictional report →05 · Repeat
Choose the required cadence, then budget for it.
Portfolio save and monitoring enrollment are separate decisions. Explicitly enable daily or weekly monitoring on a current paid plan. New entries enrolled from the saved batch use its completed check as the baseline, avoiding a duplicate initial scan. Existing entries retain their cadence and paused state.
Budget conservatively for up to five weekly or 31 daily passes in a calendar month. For 100 customers, 100 initial checks plus five weekly passes means 600 first-month screenings; 31 daily passes means 3,200. New customers and other live checks share the allowance. Free evaluation does not include recurring monitoring. Exhausted quota can stop checks, with no automatic overage charges.
The website calculator illustrates four weekly or 30 daily passes. Before enabling a schedule, use the app’s saved-draft allowance review: it compares the full conservative recurring forecast with screenings already used or reserved in the current UTC calendar month. Keep the cadence your team requires and allow headroom for other work.
After the next scheduled run, inspect the last successful check and any changed candidates. An active configuration does not prove a run completed. Assign an owner to failed or quota-blocked work; earlier review history supplies context for a new human decision. Measure the reviewer’s return separately from unattended scans.