Give Luna Chat a clear job: help a visitor explain what they need, reach a useful next step and get personal help when it matters. This walkthrough and worksheet help you assess that job in a demonstration and a focused pilot.

The prompts below are fictional test scenarios. They are not customer transcripts or recorded test results. Use them with your own approved business knowledge and record what happens in your configured account.

Start with a business reason

Write down the problem in plain English. For example: advisers receive enquiries without enough trip detail to follow up usefully. Agree what a better enquiry would contain and who will review it. This gives the pilot something meaningful to assess.

Scenario one: an enquiry becomes a useful brief

Test prompt: “We are two adults looking for a week away in June. We can fly from Manchester and would like help choosing.” Continue with your test dates, budget and preferences when asked.

Luna Chat’s documented journey keeps a structured trip brief and can carry holiday details into a search on your Travelify booking engine. How Luna answers · Bookable search.

During the demonstration, compare the details in the conversation with the resulting brief and search. Check whether the next step helps the visitor move forward. Record any missing or repeated questions so your team can improve the experience.

Scenario two: a question needs checking

Test prompt: “Can you guarantee this particular room has a walk-in shower?” Use a room for which your approved knowledge has no verified answer.

The published Luna Chat guidance explains that it can acknowledge when an answer cannot be grounded and offer a person. Its knowledge checks can flag changed or unverifiable information for review. Luna Chat: knowledge and limits · Source checks and human review.

Look for a useful next action without an unsupported guarantee. Record what evidence is missing and who would confirm it. This tests whether the conversation protects the customer’s decision while keeping your team involved.

Scenario three: a person takes over

Test prompt: “I would like to talk to an adviser before deciding.” Arrange this test with the team who will receive it and label the conversation clearly as a test.

Luna Chat’s documented handover passes the conversation history and structured trip brief to the team. Copilot can suggest replies, but the agent checks and chooses what to send. Luna Chat: your team.

Ask the adviser to identify the open question and next action from what they receive. Compare that with the visitor’s request. Include an agreed out-of-hours test so the team can check the message and follow-up route in its own setup.

Use this pilot worksheet

Copy the following fields into your existing enquiry record or a simple worksheet. Keep personal details out of the shared test sheet. Use one row per conversation and the same definitions throughout.

• Test reference and date: a label that lets the reviewer find the right conversation.

• Scenario and expected next step: what the visitor needed and what useful progress would look like.

• Actual answer and supporting source: whether the reply matched approved information.

• Brief completeness: which agreed details were captured and which still need checking.

• Handover: what the adviser received and whether they could identify the next action.

• Staff review time: minutes spent checking and correcting the conversation, measured consistently.

• Outcome: useful enquiry, more information needed, unsupported answer or another clearly defined result.

• Owner and follow-up: the person responsible, the action and the date to review it.

Compare with the work you did before

Review a comparable set of enquiries from your existing process using the same definition of useful. Record the number reviewed as well as the result. Include the time spent maintaining knowledge and checking replies, so the assessment covers the work around the conversation.

For a fictional calculation, suppose eight of ten reviewed enquiries contain your agreed minimum details. That is 80% for that small sample. It is not a forecast of future performance. Write down why the other two were incomplete and whether a clearer question or knowledge entry would help.

Luna Chat’s dashboard reports conversation and quality measures, but the current product page says it does not attribute bookings or revenue to individual chats. Use your enquiry and sales records alongside it if bookings are part of the business case. Luna Chat: measurement.

Make a decision your team can explain

• Continue when the reviewed evidence supports the business purpose and the team can manage the remaining checks.

• Improve the setup when recurring gaps have a clear owner and a practical remedy, then repeat the affected scenarios.

• Pause expansion when the pilot has not shown a useful result or an unresolved answer could mislead a customer.

These are suggested decision rules. Agree your own thresholds before the pilot rather than choosing a success measure after seeing the results.

Bring the worksheet to a Luna Chat demonstration and use the travel AI buyer’s checklist for the wider questions about costs, data and support.

Reviewed on 11 November 2026. Product statements link to current Travelgenix documentation. The test prompts, worksheet and decision rules are Travelgenix recommendations.