Where a Foundry guardrail control inspects content: user input, output, tool call or tool response, the last two being preview features for agents only.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Guardrail intervention point in context, with comparison tables and the common traps.
Terms in this definition
- WHERE
Limits a SELECT, UPDATE or DELETE to just the rows meeting a condition. Omit it, and the statement hits every row.
- Foundry guardrail
In Microsoft Foundry, a named set of controls, each pairing a risk with an intervention point and an action, that classifies prompts and completions for harms like hate and fairness, sexual, violence and self-harm, then blocks or annotates them. Formerly content filters; an agent's guardrail entirely replaces its model deployment's.
- CONTROL
Granting this on a securable gives all other permissions on it too, making it the most powerful SQL permission. At database scope that includes UNMASK and ALTER ANY MASK. Warehouse access through the Admin, Member or Contributor workspace roles carries it.
- Chat message roles
Labels on chat messages: instructions go under system, the person's input under user, the model's previous answers under assistant, and results returned by a called tool under tool (or function).
- LAST
Restricted to visual calculations, this DAX function reads a value from the final position along one of the visual's axes. Writing INDEX(-1) gives the same result.
- Document Intelligence add-on capabilities
Extra Document Intelligence options switched on with the features query parameter: barcodes, formulas, keyValuePairs, languages, ocrHighResolution, queryFields and styleFont.