DocsGetting startedInbox, testing & handoff

One inbox, with humans in control

DMtoLead centralizes connected channels so the AI can handle repeatable work and your team can step into the conversations where judgement matters.

Test before you go live

Use the AI Assistant's review and test flow with its current Training, rules and connected knowledge. Test a normal customer question, missing information and a sensitive edge case before sending real traffic to it.

  1. 1

    Try a real customer question

    Use the wording customers actually use, not only ideal prompts.

  2. 2

    Check its sources and boundaries

    Confirm it uses the correct product, calendar or knowledge source and does not invent missing facts.

  3. 3

    Improve the rule, then test again

    Fix the underlying instruction or knowledge entry instead of adding a one-off workaround that can harm later conversations.

Work from one inbox

Review and filter conversations from Instagram, WhatsApp, Messenger, Telegram and your web chat in one place. Before replying, teammates can see the customer's message history and the context the AI already used.

Hand off with context

Tell your assistant when to ask for help: complaints, order changes, price exceptions, sensitive matters and questions it cannot verify are good boundaries. It can mark a conversation for human help, and a teammate can continue with the history intact.

Manual replies are free
A teammate can take over a conversation in manual mode. Human-written replies do not use credits; resume the assistant only when the case is ready for it again.

Give people the right access

Invite your team and give each person the access appropriate to their job. Keep account, billing and assistant configuration with administrators, while agents focus on responding to conversations.