Plain meaning
A predictive model estimates a value or category. A generative model produces content. A RAG application supplies retrieved material to support generation. An agent can select or execute actions through tools. These descriptions can overlap within one application.
Three practical examples
A predictive model flags equipment likely to need maintenance. Ask about missed failures, false alarms, operating conditions, and how maintenance decisions are made.
A generative assistant summarizes a deviation. Ask whether it invents facts, omits important qualifications, or merges separate events.
An agent creates follow-up tasks. In addition to content accuracy, ask whether it selects the correct destination, uses authorized parameters, avoids duplicate actions, and recovers safely from interruption.
What to record
Describe the actual capability rather than labeling everything “AI.” Identify whether it learns from deployment data, whether changes are controlled, whether output is repeatable enough for the use, and whether it can alter records or workflows.
Completed classification statement: “The application combines retrieval and generative summarization. It reads an approved document set and produces draft text. Tool execution and write access to controlled records are disabled.”
What changes for this type of AI?
Related work: testing, supplier assessment, records, change control, monitoring.
Exercise
A RAG assistant can also send its answer into an eQMS record. Should it be assessed only as a search tool?
Show suggested answer
Include the write action and record workflow. The additional capability changes the boundary even if most users think of the application as search.
Expected records
An AI capability and permission inventory, with the system boundary and unsupported uses recorded. Adapt record names and detail to the applicable procedures; these are practitioner examples, not a universal mandatory document list.
Relevant regulations and guidance
Regulations set requirements within their scope; guidance offers recommendations or explains applicable expectations; standards depend on adoption or incorporation. Voluntary frameworks do not establish compliance. The lesson's controls, test designs and example records are practitioner suggestions unless explicitly identified otherwise.
Voluntary framework · National Institute of Standards and Technology · NIST AI 100-1 · January 2023
Scope: United States · Cross-sector AI governance and risk management, including generative AI.
Relevance to this lesson: Provides a voluntary basis for describing AI context and capabilities.
Publication status: Final / published. Source checked 21 September 2026.
Publication identity, scope and status checked against the linked official source. This is not an independent audit of every retained guide statement.
Source details and applicabilityVoluntary framework · National Institute of Standards and Technology · NIST AI 600-1 · July 2024
Scope: United States · Cross-sector AI governance and risk management, including generative AI.
Relevance to this lesson: Provides generative-AI risk context; it does not classify every AI application as a regulated medical device.
Publication status: Final / published. Source checked 21 September 2026.
Publication identity, scope and status checked against the linked official source. This is not an independent audit of every retained guide statement.
Source details and applicability