A two-question framework for deciding what work AI agents can take on

An interviewer cautions that an AI agent can sound convincing while getting a job wrong. At Salesforce Dreamforce, they discussed Mick Costigan’s framework for judging what work agents can take on: How much training data is available, and can the result be verified? The interviewer says the conversation also covered entry-level jobs, education, personal AI assistants and Slack as a place for humans and agents to work together.