What should we decide before ordering?
What decision is the AI allowed to influence?
AI knowledge assistants applies this principle to permitted documents, chunk and index pipeline and user permission check. Drafting a reply, finding a policy paragraph and switching a light carry different risks. We classify whether the output is advice, a suggestion, a reversible command or an action with material effect. The delivery timeline surveys permitted documents before design approval, then tests user permission check during commissioning. For this service, the scope records how answer with citations connects to human escalation, then verifies that relationship during commissioning or recovery testing.
AI knowledge assistants applies this principle to index and retrieval config, citation evaluation and document access owner. Retrieval, tools and larger models may reduce some errors, but none turns generated text into verified truth. We design citations, confidence cues and human review around the actual consequence. Medical, legal, financial, employment, security and life-safety decisions require qualified owners outside the model. For this service, the scope records how hosted model provider connects to index and retrieval config, then verifies that relationship during commissioning or recovery testing.
Where may data travel and remain?
AI knowledge assistants applies this principle to corpus quality, permission inheritance and citation requirement. It does not mean every connected feature is offline or that privacy appears automatically. Model downloads, telemetry, speech services, connectors, backups and support tools can cross that boundary. For this service, the scope records how update frequency connects to escalation owner, then verifies that relationship during commissioning or recovery testing.
AI knowledge assistants applies this principle to answer lacks evidence, withhold confident claim and show missing source. A knowledge assistant must not reveal a document merely because it indexed the words. Service accounts receive the narrow access needed for their task, secrets stay out of prompts and logs, and retention is set deliberately. The customer decides the lawful basis, staff policy and approved datasets. For this service, the scope records how route to document owner connects to add case to evaluation, then verifies that relationship during commissioning or recovery testing.
How is accuracy tested after launch?
AI knowledge assistants applies this principle to known-answer question, missing-answer question and revoked document. Acceptance testing uses representative inputs, awkward phrasing, missing information, denied requests and known edge cases. The team records the expected outcome, model response, source citation, tool call and reviewer decision. Maintenance repeats the known-answer question and missing-answer question checks after material changes, updates or reported faults. For this service, the scope records how conflicting sources connects to citation opening test, then verifies that relationship during commissioning or recovery testing.
AI knowledge assistants applies this principle to index and retrieval config, citation evaluation and document access owner. Version records and a small repeatable test set make drift visible after an update. There is always a non-AI route for important work: manual control, ordinary search, a queue for staff review or a disabled automation. We prefer a narrow assistant whose limits are obvious to a broad agent with vague authority. A lower-complexity alternative remains valid when hosted model provider can be handled by the documented manual or existing-system route. For this service, the scope records how hosted model provider connects to index and retrieval config, then verifies that relationship during commissioning or recovery testing.
