Contacts

A client card with a timeline, identity documents, consents, tags, duplicates and CSV import.

Available Route in Desk: /contacts

Purpose

A database of your clients with the full history of interactions. It is also where you decide whether something is a duplicate or a new person.

Timeline: calls, messages, documents, deals in one feed.
Identity documents (OCR). You upload 1 to 3 photos of a passport or ID, the system reads them through Claude, checks the MRZ and check digits and transliterates the name per the Ukrainian KMU No. 55 standard. Low-confidence fields are highlighted for manual review (threshold 0.85). Numbers and series are stored encrypted, every access to them is written to the audit log, and files are served through links valid for 15 minutes. Without an Anthropic key a demo mode works on invented data, clearly labelled.
Consents (marketing and others), tags, duplicates by phone.

This is step 4 of your quick start. The service is ready, the screen is not.

Upload a file (up to 5000 rows).
Map the columns to contact fields: fullName, firstName, lastName, phone, phone2, email, email2, telegram, viber, tags, notes, type, source, locale, consentMarketing, ownerEmail, utm_source, utm_medium, utm_campaign. Mark columns you do not need as ignore.
Review the preview before writing: for each row you see what will be created, updated, skipped or flagged as an error.
Choose the duplicate mode: update the existing contact or skip. Optionally add tags to the whole batch and a default owner (or an ownerEmail column).
Run the import. An error in one row does not cancel the others, and there is a per-row report at the end.

Phones and emails are normalized before duplicates are compared: a different spelling of the same number does not create two clients.