Page 350 - CyberSecurityEssentialsEnglish

Component What you will produce

The tool, the purpose, the accepted and rejected outputs, and the
AI usage log
verification.

TABLE 7-39: ASSESSMENT CRITERIA FOR THE EVIDENCE MAP

Criterion What earns a pass

Clarity of the question It is bounded, researchable, and states what it does not cover.

Repeatability of the
Another person could re-run it and reach comparable results.
search

Consistency of
Every exclusion is tied to a written criterion.
screening

Traceability of the
Every finding carries a page, a table, or a section.
evidence

Discipline of the
No claim exceeds the limits of the sample or the design.
synthesis

Soundness of the
Every reference exists, matches, and supports its sentence.
references

Transparency of AI use The log states what the tool did and how it was verified.

◇ Case study: from a marketing promise to a conditional finding
This composite teaching case is provided for explanation and does not describe a specific
event.
Background: an organization wants to know the effect of a chatbot on service time for
beneficiaries. Problem: vendor reports promise general results. Decision: build a map
from applied studies and operational reports. Result: it emerged that any reduction in
time depends on the type of request, the rate of escalation to a human, and the quality
of the knowledge base. Lesson: the right question was not “is the bot useful?” but “for
which types of request, and under what conditions, does it raise efficiency without

lowering quality?”

? Question to consider: can a stranger follow your chain?
Could a reader who took no part in the project start from a key sentence in your
summary and reach, within a few minutes, the source, the location, the method, and the

limits?

● Remember: Section summary
• Key idea: An evidence map gathers the question, the search, the screening, the

349