Why I Deliberately Cut My AI Usage
For a while, I used AI for everything I could. AI made a lot of my work faster and easier.
Then, I realized I was getting dumber.
I got worse at brainstorming. Why go for a walk and slowly think of ideas when an LLM can generate infinite lists of ideas? I got worse at writing. Why stumble through multiple drafts and edit when AI can generate infinite content?
So I significantly cut my use of AI. I want my ideas to come from my experiences. I want my writing to reflect my actual beliefs. And I want my brain to “struggle” to figure things out.
The Microwave Problem: When AI Makes Sales Feel Efficient But Taste Wrong
Derek H nailed the risk of AI overuse specific to sales in a post that caught my attention:
You can cook a steak in 3 minutes in a microwave. It’s efficient. It’s fast. It’s “productive.” But you’d never serve it to your most important client.
In tech sales right now, we are obsessed with the microwave. We’re using AI to scale our discovery, automate our research, and optimize our outreach. On paper, the productivity metrics look solid. We’re doing more than ever.
But there’s a catch. When you get in the room or on the Zoom with a customer, they can taste the microwave. They can tell when your discovery questions were pulled from a generic prompt rather than genuine curiosity. They can feel the rubber texture of a pitch built by an algorithm that doesn’t understand their internal politics or the stakes of their role.
AI can prep the ingredients. But it cannot provide the sear.
This microwave analogy captures something critical about AI in sales. For complex evaluations, people still want to talk to people. For simple evaluations, people will gravitate to the convenience of AI and automation.
Products that can be sold with a “script” will almost certainly be automated by AI. But products that require complex reasoning to evaluate — the transformational decisions that enterprise sellers handle — require human judgment, adaptation, and relationship-building that AI cannot replicate.
The Right Way to Think About AI in Sales
I remain optimistic about the use of AI in sales. I’m heavily involved in rolling out new agentic workflows for our sales org, constantly evaluating new tools, and have championed the purchase of multiple AI solutions.
The key is understanding when AI enhances your capabilities versus when it replaces your thinking. AI excels at accelerating research, generating initial drafts, and organizing information. But it cannot understand the nuanced politics of your prospect’s organization, read between the lines in a discovery call, or adapt your approach based on subtle buyer signals.
When I use AI with my team, I emphasize that you never accept AI output as facts. Every AI-generated insight needs validation. Every AI-drafted message needs personalization based on your actual knowledge of the prospect. Every AI-suggested approach needs your judgment about whether it fits the specific situation.
The sellers who succeed with AI use it to accelerate their preparation, not replace their thinking. They use AI to generate options, then apply their expertise to select and customize the best approach. This is why “default to AI” is bad advice for enterprise sellers — it removes the human judgment that creates differentiation.
How Do You Know If You’re Too Dependent on AI?
Sellers who rely on AI for everything show three warning signs: they can generate good outputs but don’t understand them, they can create accurate plans but can’t extract the most relevant pieces, and they can prompt LLMs but struggle in actual customer conversations.
Here’s what AI over-dependency looks like in practice:
You can generate a good call plan, but don’t understand it. AI creates a comprehensive discovery framework, but when the prospect takes the conversation in an unexpected direction, you can’t adapt because you don’t understand the underlying logic of your own plan.
You can create accurate account plans, but can’t extract the most relevant pieces. AI researches the account thoroughly, but when you’re in front of the buyer, you can’t quickly surface the most compelling insights because you didn’t internalize the research — you just copied it.
You can prompt LLMs effectively, but can’t talk to customers. You’ve mastered prompt engineering, but your discovery calls feel scripted because you haven’t developed the conversational skills to dig deeper when prospects give you gold nuggets.
The most dangerous part? These sellers often look productive on paper. They’re generating more outreach, creating more account plans, and producing more call summaries than ever before. But their conversion rates tell a different story.
The Bottom Line: Enhancement vs. Replacement
If the only output you can create as a seller comes from AI, you’ve already been replaced. Payroll just hasn’t quite caught up yet.
The future belongs to sellers who use AI to accelerate their preparation while maintaining the human skills that create trust, adapt to complexity, and build relationships. Use AI to research faster, draft better, and organize more effectively. But never let it replace your ability to think, adapt, and connect with buyers.
Your prospects can taste the microwave. Make sure you’re still providing the sear.
