This one’s important. Our stance, honestly, is “not exactly.”
There are three schools of thought here:
Those who believe in ‘role replacement’ capabilities for a single AI agent (e.g. “A Marketing Manager Agent”)
The ‘role enhancers!’ framing of agents (e.g. “research assistant”)
The more honest framing, that agents are great at specific task/workflow execution and do way more when they work together.
Both 2 and 3 are, in our opinion, ‘true’ depending on the use-case. The most common use-case is certainly just using agentic capabilities like thinking and tool calling within a chat interface to augment human work. Things like “I need to book dinner on Friday, can you find a table for 2 around 7pm and do it?” would be a good example here.
For leaders, entrepreneurs, teams, people, etc. trying to fully automate functional roles within their org (like a marketing manager) it’s important to remember that most jobs actually involve loads of interconnected tasks and operations with varying levels of cognitive, creative and interpersonal skills involved. It’s really best NOT to frame AI Agents as a full-on replacement of human job functions, yet.
Maybe in the near future.
But for now, job function replacement is being achieved more realistically through agentic mesh implementations, where manager/operator agents have specific functionality and work with other specialized agents (think: “Instagram Post & React Agent") to handle various needs.
Instagram poster isn’t really a job you hire, but it might be an AI agent you build.
