Skip to main content

What data does chat use to respond?

NEXT AI chat only uses (a) the prompt you type and (b) the specific customer-evidence supporting your prompt.

Inputs to the chat​

  • Your prompt.
  • The memories you choose as context (e.g. all memories in your teamspace, a search/filter subset, memories from selected records, or just one memory). You fully control that scope.

Where those memories come from​

  • Call records you import from tools like Zoom/Google Meet/Gong are processed, key customer moments are identified or clipped, interpreted, and classified, with traceability back to the source.
  • Support tickets & conversations (e.g., Zendesk, Intercom, Front, Zoho Desk).
  • Surveys/VoC (e.g., Qualtrics, Typeform, Mopinion, GetFeedback) — each response becomes a memory with scores/metadata plus a source link.
  • App reviews (Apple App Store, Google Play) — imported as memories with rating, version, and a source link.
  • Communities/other sources (e.g., Khoros) and CRM context (Salesforce accounts/opportunities) to anchor work and automations.

How answers are grounded​

  • NEXT AI generates answers only over the selected memories, and the UI links back to the exact source artifact (record/ticket/review), including reels/clips for video so you can verify.

Chats are private to the creator by default (you can explicitly share them with teammates).