Migrate from Atomic CRM to Nutshell

Atomic CRM Contact records map directly to Nutshell Person records. The UUID primary key from Supabase is discarded; we use email as the dedupe key during Nutshell import. Standard fields (first name, last name, email, phone, address) transfer without transformation. Any custom fields added as Supabase columns on the contacts table migrate as Nutshell custom fields, which we create in the destination account before the Person import phase begins.

Atomic CRM Company records map to Nutshell Company with a one-to-one correspondence on name, domain, and address fields. We import Companies first so that the Company-to-Person association is satisfied before Contacts are inserted, preventing orphaned Person records. Nutshell's Company records do not have a native industry classification enumeration, so any industry custom field from Atomic CRM maps as a text or picklist custom field in Nutshell.

Atomic CRM Deal records map to Nutshell Deal objects. The dealstage value originates as a TypeScript enum in App.tsx — we extract the enum values during scoping and map each to a corresponding Nutshell pipeline stage name during schema design. Deal value, expected close date, and owner assignment transfer directly. Deals with no associated Company link to a Person instead. Nutshell supports multiple pipelines, which we configure to match the number of deal categories present in the source enum.

Because deal categories are props passed to the CRM root component in App.tsx rather than stored in the database, the migration extracts enum values from that file during discovery. Each enum entry becomes a Nutshell pipeline stage. We preserve the original enum key name in a custom field original_stage_key__c on each Deal for audit. The customer must share App.tsx or equivalent during scoping; without it, we cannot enumerate all stage values and some Deals may arrive in a catch-all stage.

Atomic CRM Notes linked to Contacts or Deals migrate to Nutshell Notes attached to the corresponding Person or Deal record. Note body text, author reference, and creation timestamp transfer. Rich text formatting in Nutshell Notes is more limited than free-form text in Supabase, so long-form notes with embedded HTML may lose formatting; we strip to plain text as a fallback and flag any affected records during reconciliation.

Atomic CRM Task records (with task type derived from TypeScript props) map to Nutshell Tasks. Task type values (Call, Email, Meeting) are extracted from App.tsx during scoping and mapped to Nutshell task type categories. Status, due date, assignee (resolved by email match to Nutshell User), and creation timestamp transfer. Completed-at timestamp migrates as a custom field since Nutshell Tasks carry status but not a separate completion timestamp.

Custom fields in Atomic CRM exist as Supabase table columns added via Supabase Studio — there is no in-app custom field UI. During scoping, we inspect the full Postgres schema across all tables (contacts, companies, deals, notes, tasks) to surface every column, including any added without documentation. Each discovered column maps to a Nutshell custom field of equivalent type (text, number, date, picklist). We create the Nutshell custom fields before the relevant record import phase begins so that all fields are available during data mapping.

Atomic CRM User records are Supabase Auth UUIDs. We map source Users to Nutshell Users by email address rather than UUID, since UUIDs from Supabase Auth have no meaning in Nutshell's user directory. Any Supabase Auth UUID referenced on a record but without a matching Nutshell User email is held in a reconciliation queue for the customer's admin to provision before record import resumes.

Atomic CRM's 200+ React hooks and its FakeRest data generator logic (src/components/atomic-crm/providers/fakerest/dataGenerator) implement custom automation behavior that has no direct Nutshell equivalent. We do not migrate hooks or data generator code. We document every identified hook and generator file with a description of its behavior and a recommended Nutshell workaround (workflow, Zapier/Make integration, or manual process) in the written automation inventory delivered at cutover.