What happens next
Your AI hands you HubSpot’s own sign-in link. You approve read access — Green Trees can look, never touch. Nothing in HubSpot changes, during or after.
Your AI pulls a sample — companies, contacts, deals, with their fields — so you both see what’s actually in there.
It proposes where every field lands and asks before anything moves. Odd custom fields, duplicate records, half-filled columns — decided in conversation, not a spreadsheet.
Records import in resumable batches, each filed through the same dedupe every record gets: existing entries are skipped, never overwritten. Every batch leaves a receipt.
HubSpot keeps running untouched while you settle in. Re-running the import is safe — it only picks up what’s new.
Answers you’ll want first
What moves?
Companies and contacts today; deals are next on the list. Notes and custom fields ride along with the mapping you approve.
Can it break my HubSpot?
No. The connection is read-only by construction — there is no write path back.
What about duplicates?
Import files through the same skip-never-overwrite dedupe as everything else. A record you already have is left exactly as it was.
Salesforce?
Same conversation, same read-only doctrine — tell your AI which one you’re leaving and it takes it from there. See how the three compare.