The short answer

Communications AI can organize conversation context and prepare suggestions, but it does not replace source records or user judgment. AI suggests; a person reviews, approves and sends or applies an action.

Availability: Available to accounts with access to this workspace or feature.

Step-by-step

  1. Review the conversation summary

    Check the customer request, important facts, previous commitments, unanswered questions, latest outcome, recommended next action and overall relationship history. Refresh when newer customer activity makes a summary stale.

  2. Draft and refine a reply

    For supported SMS, email, form and booking conversations, choose Draft with AI, then optionally make it Shorter, Friendlier, Professional or translate it. The assistant may use server-grounded company profile, operating hours, services and prices.

  3. Verify sources and wording

    Open the source disclosures, compare the draft with the original customer message and correct unsupported claims. Use draft only places text in the composer; AI never sends automatically.

  4. Interpret lead qualification

    Treat the lead score as an estimate. Review confidence, evidence and structured criteria such as requested service, location, budget, timing, business size, buying intent, urgency and decision-maker status.

  5. Configure and preview score rules

    Company-defined qualification rules can change score effects. Preview the effect before saving, and do not treat missing or low-confidence facts as confirmed customer information.

  6. Review next-best actions

    Recommendations such as call, send packages, request documents, offer times, schedule follow-up, assign or resolve include reasons, capability checks and an effect preview. Confirm an available action yourself; unavailable actions produce no side effects.

Was this guide helpful?

Send feedback about this guide