Skip to Content

Three examples to test

Configuration patterns to adapt with support, not customer testimonials or guaranteed results. For each, verify the collected data, response and next step before launch.

Restaurant: information and table enquiries

Goal: answer recurring questions and collect a request without constantly interrupting staff.

  • Prepare: opening hours, services, terms and approved menu information. Refer sensitive dietary questions to staff, without unverified reassurance.
  • Collect: name, contact details, day, time, party size and necessary notes.
  • Expected result: a complete request available through the agreed channel. Confirm the table only after a successful result from the connected system.
  • Exception: when availability cannot be checked, capture the preference and explain that the team must confirm.

I can take your request for four people at 8 pm. The restaurant still needs to confirm the booking.

  1. Ask about opening hours and request a table.
  2. Change the party size during the call.
  3. Check that the final data is correct and no unverified availability was promised.

Gym: information and trial enquiries

Goal: capture interest in a class or trial and give the team what it needs to follow up.

  • Prepare: approved classes, hours, prices and trial terms.
  • Collect: name, contact details, class of interest and preferred availability; avoid unnecessary health questions.
  • Expected result: recorded interest and preferences. An appointment is confirmed only when the connected calendar accepts it.
  • Exception: a promotion missing from the sources or an out-of-scope request needs team review.

You are interested in trying the evening class. I will collect your availability and contact details to arrange the next step.

  1. Ask for a listed price and an unlisted promotion.
  2. Correct your contact details before ending the call.
  3. Check the class, preference and contact details in the call result.

Call center: qualification and callbacks

Goal: test an outbound AI campaign with a small test list, collecting interest and next steps. Connections to human operators must be agreed separately: they do not imply native operator workstations in the platform.

  • Prepare: usable contacts, goal, custom fields, priority, calling hours, time zone and callback rules.
  • Collect: interest, agreed information and any callback preference.
  • Expected result: calls, outcomes and data linked to the campaign; a requested callback must follow the configured rules.
  • Exception: for no answer, check the expected behaviour; for a do-not-contact request, stop follow-up for that contact through the agreed procedure.
  1. Test an interested contact, a callback request and a no-answer case.
  2. Check calling times, attempts, outcomes and saved data.
  3. Validate how do-not-contact requests are handled before expanding the campaign.

From an example to your agent

Choose an example and note what the agent should do, which data it can use and what should happen if it cannot complete the request. Setup and tests must be adapted to your service.