Why does the bot behave oddly when I run many tests in a row?

If you test the bot many times from the same number or the same session, it may carry over previous context and produce strange or inconsistent replies. That does not always mean the flow is misconfigured.

What usually happens

During repeated testing, the bot may:

• Skip questions it asked before
• Remember a previous branch or preference
• Mix languages or earlier responses
• Assume information from an earlier conversation

This happens because the session keeps some context so a real customer conversation can continue naturally.

How to fix it

The fastest solution is to test from a number or account you have not used with that channel before: another phone, a colleague’s, or a different Telegram account. The bot will treat it as a brand-new customer and start the flow from scratch.

If you need to repeat the test from the same number with a clean session, write to us and we will help.

When to do this

This is especially helpful when testing:

• the initial welcome flow
• branch selection
• language switching
• staff selection
• brand-new bookings

Good testing practices

• Run one scenario at a time.
• Use a fresh number or account before switching to a different use case.
• Do not judge the bot based on a session that is heavily “contaminated” by earlier tests.
• If the issue still happens in a clean session, then it may be a real configuration problem.

When to contact support

Reach out to support if:

• The issue repeats from a fresh number or account too
• The bot ignores correct availability or branch data
• The problem happens with real customers too, not only in your own testing

In short: if the bot seems confused after many tests, repeat the test from a fresh number or account. It is one of the most common causes of misleading results during testing.