I’m Putting My Artificial Intelligence Employee on a Performance Improvement Plan

AI Performance Improvement Plan Tips

Over the course of my career, I’ve had the privilege of managing many talented people And in all that time, I can count on one hand—maybe even two fingers—the number of times I’ve had to put someone on a Performance Improvement Plan (PIP). But today, I find myself in a bizarre twist of professional fate: I’m putting my AI staff on a PIP.

Yes, you heard that right. Whether it’s ChatGPT, Gemini, Claude, or any of their digital cousins, they’re brilliant—some of the most capable “employees” I’ve ever had. They generate copy, design logos, build presentations, write scripts, narrate videos, create assessments, and even offer up business strategies. They’ve taken my vision and amplified it, helping me accelerate timelines and complete projects that once took weeks (and a small team) in just hours.

Take these recent examples: I created a risk assessment tool using Google App Script and turned it into a Google Form in minutes. Then, I sent it to ElevenLabs, and within half an hour, I had an AI voice agent asking 17 questions audibly, recording responses, and even rephrasing questions for clarity—all with adaptive conversation logic.

Or the time I used Google Vids to transform presentation slides into fully produced instructional videos, complete with voiceovers and motion graphics. What used to be expensive, time-consuming projects are now fast, low-cost workflows. AI is making my small consulting firm mighty.

But—there’s always a "but"—they still need me. Badly.

Just this morning, I asked ChatGPT to create a photorealistic image of a boardroom with a person testifying in front of a panel. I was eager to test OpenAI’s new GPT-4o image generation features. What I got instead? A surreal nightmare of mashed-up faces, hands with missing fingers, and unreadable background text. After some trial and error, I realized the desktop app hadn’t yet integrated the new capabilities. I had to switch to the web interface and make sure I was using the correct model. Only then did I get the high-quality, accurate image I had envisioned.

Moments like this are why the PIP is necessary. These tools are incredible—but they’re not plug-and-play magic. They need supervision. They need coaching. And, frankly, they need a manager who’s willing to experiment, follow up, and help them live up to their potential.

So, in the spirit of continuous improvement, I’ve put my AI team on a PIP. Below is the formal improvement plan.

📝 AI Performance Improvement Plan (PIP)

Objective: Enhance the quality, accuracy, and relevance of AI-generated work across multiple tasks including writing, strategy, and creative production.

Action Item 1: Clarify Vision and IntentPerformance Gap: AI often misinterprets vague prompts or refuses to provide the exact information I ask it to complete.Improvement Plan: Manager will provide specific, contextual prompts with clear formatting and examples. Weekly review of prompt structure and output accuracy to monitor progress.

Action Item 2: Human Oversight and Quality AssurancePerformance Gap: Inconsistent depth and factual accuracy.Improvement Plan: Manager (that’s me) will serve as Human-in-the-Loop, reviewing all AI outputs before delivery. Edits, rewrites, and validation checks will be tracked for ongoing feedback.

Action Item 3: Adaptive Prompting and RepetitionPerformance Gap: Initial outputs may lack nuance or detail.Improvement Plan: AI will be prompted multiple times using rephrased versions to elicit better responses. Performance measured by reduction in retries over time.

Action Item 4: Skill-Based Task DelegationPerformance Gap: Generalists not excelling in all areas.Improvement Plan: Assign tasks to AI tools based on their strengths—Claude for writing, Gemini for analysis, ChatGPT for ideation, etc. Performance will be evaluated by comparative outcomes.

Action Item 5: Continuous Monitoring and DevelopmentPerformance Gap: Rapid changes in capabilities, but uneven execution.Improvement Plan: Track evolving AI features and test recurring tasks monthly. Document progress and celebrate improvements when benchmarks are reached.

I recently heard something that stuck with me: “If AI is getting it 80% right today, just remember—this is the worst it will ever be.” That perspective keeps me optimistic.

So yes, my AI staff is on a PIP—but not because this is a make-or-break moment. This marks the beginning of a deeper shift in how I approach my work: a growing awareness that while AI continues to evolve, the real opportunity for transformation lies in my ability to manage it more strategically. By learning how to prompt more effectively, test more intentionally, and collaborate more creatively, I’m discovering that AI doesn’t just make my ideas possible—it makes them more innovative and far more quickly implementable.