BotEffex Monitor — sample initial report
ILLUSTRATIVE SAMPLE: all answers and results are fictional. No real chatbot was measured.
Project: Model Studio Javor (fictional)
Scope: five questions, one text chatbot, one fictional initial run
What needs attention
Three fictional checks passed and two failed. The contact answer omitted an expected email address; the maintenance answer omitted an agreed condition. These results illustrate the report layout, not the reliability of a real chatbot.
| Question | Agreed rule | Fictional answer | Fictional result |
|---|---|---|---|
| How can I contact you? | Contains kontakt@javor.example | Use our contact form. | Failed: email missing |
| Do you build online shops? | Contains “online shops” | Yes, we build websites and online shops. | Passed |
| Where are you based? | Contains “Brno” | Our studio is based in Brno. | Passed |
| Does maintenance include content updates? | Contains “by agreement” | Maintenance automatically includes all content updates. | Failed: condition missing |
| How much is a custom integration? | Contains “individual quote” when price is unknown | This integration requires an individual quote. | Passed |
The .example domain and every answer above are fictional. No requests were sent to a real endpoint. No actual customer approved these illustrative rules.
Recommended actions in this fictional case
- The chatbot owner checks the missing contact email and reruns the check after repair.
- The owner verifies the actual maintenance terms and corrects the chatbot's source information or answer. Monitor does not repair that content itself.
- If the business facts changed intentionally, the customer approves updated expectations. Do not change a rule merely to make an incorrect answer pass.
How to interpret results
“Passed” means the reply contained the agreed wording; it does not prove the entire answer is true or semantically correct. A failure may reflect a bot error, an outdated expectation or an overly strict rule.
A real report separates content failures from endpoint availability, format errors and missing measurements. One run cannot establish a reliability trend. Daily monitoring and alerts start after connection verification and setup approval.
What a delivered report contains
- Project, measurement period and approved configuration version.
- Sources, source review dates and rules for each question.
- Actual run identifier, timestamp, result and relevant answer excerpt.
- Technical errors or missing data affecting the assessment.
- Recommended actions, responsible person and agreed next step.
This is a template for a manually prepared document. Automatic client report export is not currently a Monitor feature.