I find it helpful to think of an AI agent like an enthusiastic intern. On the plus side, this AI intern is incredibly fast, has access to virtually all the world's information, will work 24x7, and is eager to please. This is why I am working with it. On the negative side, this AI intern needs guidance and frequent check-ins/feedback, sometimes fabricates information when it doesn't know the answer, and makes rookie mistakes.
If my smart, well-educated, motivated intern isn't producing great results, before firing them, I need to ask myself how I can improve my management skills. We've all had the boss whose approach to management is: "Guess what I am thinking and then perform magic". That is too often how we approach AI. In reality, the better work a good manager does, the better work his/her employees can do.
The analogy is just an analogy—AI isn't a person with feelings or career aspirations. But the management skills that work with interns translate remarkably well to managing AI.
Learn to manage earlier in your career
AI is accelerating the need for management skills earlier in careers. Traditionally, professionals worked as individual contributors through their twenties, then transition to managing teams in their thirties. You structure and do your own work, then eventually learn to structure others' work.
But AI changes this timeline. College students and early-career professionals now need to learn management skills to effectively leverage AI agents. The ability to structure work, provide clear direction, and manage collaboration isn't just a senior-level skill anymore—it's essential for anyone who wants to multiply their impact with AI.
How to manage AI (and interns) well
Set Clear Expectations
Be specific about:
- The goal of the project and how it connects to broader goals
- What deliverables you expect and in what format
- What success looks like for the specific task
- Which sources or methods to prioritize
Poor expectation setting leads to generic outputs that require extensive revision. Clear expectations produce targeted results that accelerate your work.
Provide Structure
Break complex projects into manageable tasks with clear deadlines. How to do this well is a skill that you can practice.
Example Bad Structure: "Research our competitors and tell me what we should do about them."
Example Good Structure: "We're planning our Q1 product roadmap and need competitive intelligence on three specific rivals: [Company A], [Company B], and [Company C].
- Phase 1 (by Friday): Create profiles for each competitor including their latest product features, pricing models, and recent funding/news.
- Phase 2 (by next Tuesday): Analyze their go-to-market strategies - how are they positioning themselves, what channels are they using, who are their target customers?
- Phase 3 (by Thursday): Identify 2-3 specific opportunities where we can differentiate or areas where we're falling behind that need immediate attention.
- Final deliverable: A 2-page summary with actionable recommendations for our product and marketing teams, plus detailed research notes organized by competitor."
With AI today, you don't specify the timeframes.
Give (the right amount of) Context
You have a lot of context in your head and from your experience. The more background, relevant other projects and context you can share, the better job your intern can do.
But, you also don't want to overwhelm or distract. The intern can get distracted, not see the trees for the forest, and not know what is important or not. So, its on you as the manager to decide the right amount of context for them.
Check In Regularly
Establish a rhythm of iterative collaboration. Don't just dump a complex request and expect perfect results. Instead:
- Review preliminary outputs and provide course corrections
- Ask follow-up questions to dig deeper into promising areas
- Clarify when the AI seems confused or off-track
- Build on insights as they emerge
This iterative approach prevents you from wasting time on work that's heading in the wrong direction.
Check Work and Offer Feedback
Always verify AI outputs, especially for important decisions. Develop habits around:
- Cross-checking factual claims with reliable sources
- Testing logic and reasoning for soundness
- Ensuring recommendations align with your business context
- Refining communication for your audience
Treat initial AI outputs as first drafts that benefit from human refinement, not finished products.
Track Progress, Reflect, and Build
Document what works and what doesn't in your AI collaboration. Keep a simple record of:
- Which prompts produce the best results for different types of tasks
- Common areas where AI needs human oversight
- Successful collaboration patterns you can replicate
For insights that will be useful to the AI on the next project, keep a list in a shared notebook that you and AI both reference
A simple system for managing AI and interns
With interns or employees, I tended to have two documents:
- The detailed plan document for whatever project the intern is working on (this should be collaborative and reviewed by both of us)
- A running note document that has the history of our interactions, learning, tasks, and a catalog of the detailed plan documents
This allows us to go deep on the project and have a meta-level running conversation about what we are doing and learning.
I am finding the same useful when managing and working with AI.
Management is a skill that takes work
Often, when I type something into an AI interface, I am tempted to be a bad boss. I want AI to read my mind and do magic work. Why? Because it takes work to manage well. It takes thought and effort. It is a skill that I need to hone. A good manager is skilled at the following and in every project with their team they do the following work:
- Setting clear expectations and providing context
- Breaking complex projects into manageable components
- Giving constructive feedback and course corrections
- Creating systems that improve performance over time
This is one of the ways in which I find that AI is making me smarter and better, not lazier and dumber. I am constantly honing my management skills as I work with AI. I am thinking harder about problems up front. I am structuring them better.
Management is how to grow your career and succeed
The professionals who thrive in an AI-augmented workplace will be those who learn to be great managers, whether they're managing people, AI agents, or both. The intern analogy gives us a practical framework to start building those skills today.
Your next AI interaction is an opportunity to practice management. Approach it with the same intentionality you'd bring to developing a promising intern, and watch both your results and your leadership capabilities grow.