Customer Support Bot Test Suite

Problem/Feature Description

Acme Devices recently launched a new customer support bot for their online store. The bot handles a range of common interactions: greeting customers, answering questions about products and their support phone number, and returning a polite error message when something goes wrong internally. Because the bot is powered by a third-party NLP service, its exact phrasing can vary slightly from run to run — for example, it might acknowledge a question with "Sure!", "Of course!", or "Happy to help!" before giving the actual answer. Your team wants a solid automated test suite to catch regressions as the bot evolves.

The bot is a pure Python function — no external service or network call required. You'll implement a small stub that covers the four interaction types, then write pytest tests using the pytest-conversational library to verify it behaves correctly. The test suite should cover: that replies contain expected keywords, that error replies don't expose internal details, that formatted contact information (phone number) is present and well-formed, and that flexible acknowledgement phrases are all accepted as valid responses.

Output Specification

Produce the following files in your working directory:

• test_support_bot.py — a pytest test file containing the bot stub and the full test suite. The file must be runnable with pytest test_support_bot.py -v.
• results.txt — the captured output of running pytest test_support_bot.py -v, saved to this file.

All tests should pass. The bot stub and tests should be self-contained in test_support_bot.py with no external dependencies beyond pytest-conversational (install it with pip if needed before running).