Reporter Tests AI Agent to Perform Her Job

Business Insider reports that reporter Amanda Hoover built an AI agent she called "Amanda Bot", trained on her past work, and tasked it with end-to-end reporting tasks. Business Insider reports the bot conducted interviews through a voice agent, attempted to draft the story, and navigated interaction with an editor. Business Insider reports that, after the exercise, Hoover wrote that the experiment "didn't leave her worried about losing her job to AI."

Editorial analysis - technical context: Public coverage frames this as a practical test of modern AI "agent" workflows rather than single-turn LLM prompts. Industry reporting in the article highlights multi-step agent capabilities - scheduling and interviewing via voice, multi-turn drafting, and autonomous follow-up - which align with the broader move from isolated tools to chained, objective-driven systems. The article includes a direct quote from Aravind Srinivas: "With AI, you don't tell the computer what to do step-by-step anymore. Instead, you tell it what you want, you lay out your objective, your plan, and the computer reasons and executes your plan," Aravind Srinivas, cofounder and CEO of Perplexity AI, said at a recent event, Business Insider reports.

Industry context: Business Insider frames the experiment amid growing adoption of worker-facing agents. The article profiles practitioners who run multiple specialized agents, including a ClickUp growth-operations manager identified as Cabasso who reportedly gives agents fictional personalities and models a vendor-negotiations agent after an FBI hostage negotiator, Business Insider reports. This coverage fits a broader pattern where teams move from one-off automation to delegating compound tasks to agents, raising questions about reliability, oversight, and role boundaries.

For practitioners: observers should track three indicators reported and implied by this experiment, agent fidelity on interview and sourcing tasks, the extent of human editorial intervention required, and how organizations assign or restrict agent autonomy across seniority levels. The Business Insider experiment suggests agents can automate routine, structured work but that human oversight remains central for nuanced reporting and editorial judgment.