A utility must pull structured fields from scanned inspection forms and flag field photos that show an unsafe condition. Clerks currently retype the forms. Which capability should the architect propose?
Select an answer to reveal the explanation.
Short Explanation
Reading a stack of scanned forms and looking at the photos are skills, not a list of trigger phrases. You propose Foundry Tools to pull the fields and to flag unsafe images. A formula that only files the photo leaves the clerk retyping.
Full Explanation
Foundry Tools is the family to propose when the requirement is a specific perception or language skill rather than a conversation. Document extraction covers the scanned inspection forms, and an image safety check covers the unsafe-condition photos, so the clerks can stop retyping and start reviewing exceptions. Both skills already exist as tools, which is why a new conversation topic is the wrong place to start. A topic full of field-name trigger phrases never opens the scan. It only fires if someone types those words. A canvas formula that stores the photo handles filing, not reading, so the form fields and the unsafe condition stay invisible. Business terms correct how a copilot interprets the organization's words, and they do not extract a scan or classify a photo. Exam caveat: the agent that talks to the clerk can still live in Copilot Studio. It should call these tools rather than pretend a topic is an extractor. Operational check: submit one known form and one photo the safety reviewers already consider unsafe, and confirm the fields match the form and the photo is flagged.