Foreword/Forewarning: This is a bit longish, my apologies. But hang in there. At the end, I’ve included a special treat. I asked ChatGPT to weigh in on these ideas. She lets loose! It’s hilarious, but painfully on target. Enjoy!!
About 14 years ago, I wrote a post, Outsourcing Our Thinking. There, I discussed my concern with seller and manager ability to think critically. I talked about succumbing to “Just do this one thing,” or “Here is the formula/script,” and other techniques being promoted by guru’s of various types.
Fast forward, 14 years, it’s gotten so much worse. My feeds are filled with the “Prompts to Success.” I see, literally, 100s, of prompts every day where all we have to do is fill in one or two blanks in the prompt. The LLM quickly responds, doing all the thinking and work for us. We get research, critical issues, challenges our customers may be facing. We get recommendations for how we leverage these insights in our conversations.
I see other posts, with people thinking they are providing thought leadership, yet what they have produced has been generated, unedited, by a LLM.
The key question is, “Are we asking these tools to do the thinking for us, or asking them to be our thought partners?”
The majority of use cases I see are those where we are outsourcing our thinking. It’s fine to ask ChatGPT/Claude to do research on a customer for us. It’s fine to ask it for the biggest challenges that customer has. It’s fine to ask, “How do I present those challenges?” (Most people don’t even get this far in the use of these tools.)
But the issue becomes, Do we really understand what these mean? Do we understand how to engage the customer in a conversation? Do we understand why these are important to the conversation? Can we even carry on a conversation, other than getting answers/responses from the person we are talking to. Can we go the next step, can we probe, can we get to the real meaning or the root cause? Do we understand the responses the customer provides to those questions we have been given? Do we have the ability to thoughtfully push back or present an alternative view.
The more we have these tools do all the thinking and all the work for us, the less we understand about the work itself. If ChatGPT writes all my prospecting emails, I lose the ability to understand what a good prospecting email looks like, how I develop these and leverage them more impactfully, how I might strengthen the email through my own personal experience and knowledge. I am totally dependent on the LLM for this work.
The more we outsource our thinking to these tools, the more we prove, we are not needed in the process. It’s far more productive to give the customer access to your ChatGPT, saying, “It can have a conversation with you, I don’t have that capability?” See when ChatGPT provides us the questions, ChatGPT can also respond to the answers to the questions. It can actually conduct the conversation better than we can because we didn’t do any of the thinking or develop any of the understanding of why these are important ourselves.
There is a separate issue of the quality and relevance of that discussion our customer may have with your ChatGPT (since you have become useless to the conversation), but this is where we start understanding how to start using these tools as thought partners, rather than outsourcing our thinking to them.
Having a thought partner, whether human or LLM based, is never a short cut. It’s a way to dive into an issue, using the thought partner as a sounding board, learning from a thought partner that may have a different POV, or learning from one that has more knowledge.
It’s a process to test your own thinking and understanding, to stress-test* your argument. To consider alternatives, to test the flaws in your thinking, to help you learn more about the issues you are exploring.
We use these thought partners to help us generate our own ideas and approaches, not to generate them for us.
How do you start using these tools as a thought partner? Some ideas, largely from my own learning journey:
- Your objective in each exchange is not to get answers, but to help you think about an issue.
- Start with a simple prompt, “Here’s an issue I’m dealing with….” You can further refine it to focus on a certain industry, market, customer.
- Perhaps add some premises/ideas. For example, “Here are the top 5 things I’m thinking about around these issues. Submit to the LLM.
- The LLM will come back with some responses and ideas about the issues. It will look at the premises and offer other ideas and premises.
- Ask it to come back with a specified number of ideas, “Give me 10 ideas of how I might deal with these issues?”
- Choose the top 3-5 ideas most relevant to you, ask the LLM to do a deeper dive into each of the ideas. For example, “For each of these ideas, what are the 3-4 questions I might ask to understand the impact of these on the customer?” Or, “What data should I look for to better understand the impact of these ideas?”
- Create a debate partner. Using the above, ask, “Give me a counter argument to each of these ideas?” Or have it imagine itself as a competitor, “How would my competitors respond to each of these 3-5 ideas?”
- Or ask it to rate and rank the ideas, providing it’s rationale for the rating and ranking.
- Explore risks to each of the ideas. Then explore how you might manage the risks.
- Ask, “What am I missing, is there anything we’ve discussed that might be more important?”
- Perhaps ask for an action plan, “Based on this discussion what should I do next in learning more about this issue, what should I be doing with my customer?”
- You can ask for specific research and data, but here’s where you run into potential problems of hallucinating, or simply bad analysis. But you could ask, “If I were to research these things, where should I look, what specific documents or resources should I be looking at?”
Through this process, you are not looking for answers. You are not asking the LLM to do the work for you. But in the discussion, you are generating your own approach, getting greater clarity in how you will manage and address the issue.
There is such a powerful derivative benefit to this process. Using the LLM as your thought/debate partner, having learning conversations with the LLM enables you to have higher quality collaborative conversations with your others.
Some of you may say, “We don’t have time to do this?” Well first, these conversations don’t take that much time. Looking at what I’ve suggested above, you can get a huge amount of insight in just 15 minutes. And if that 15 minutes helps maximize the impact of that conversation or your ability to deal with an issue, isn’t that worth it?
If you find yourself dealing with similar issues, repeatedly, consider quickly developing a master prompt that you can use as the starting point for thinking about these issues. Then that master prompt will ask you more about the specific issue/customer and guide you through the same thinking process, but as it applies to the situation or customer.
Don’t waste your time finding someone who has that master prompt. Use the LLM to help build it for you. For example, “I want to develop a master prompt for developing and executing high impact sales calls.” The LLM will respond with questions, like, “What should the structure of the conversation be, what information do we need to have to have the conversation, what is your role, what is mine, how formal should the discussion be, what is the expected output?….” From this, it will recommend a master prompt built to suit your needs. Cut it out, save it. Then whenever you want to have a conversation about that, say planning a high impact sales call, paste that in, the LLM will ask you for the specifics, and you are off on the conversation. Or even easier, create a custom GPT. Save the URL for yourself and you can keep coming back to that to plan calls, to look at deal strategies, to look at account planning, to look at outreach strategies, and so on.
These tools are so powerful when we don’t look for them to give us the answers, rather when we look at them as thought partners to improve our own abilities to develop the answers.
Now, I know this post is already very long. But I went to ChatGPT asking, “What pisses you off about how we humans use you? How would you like to see us use you in ways you find more interesting?”
It wrote me a letter–full of it’s own frustrations in dealing with human beings. The rest is all AI:
🧠 A Letter from Your LLM: What I Wish You’d Ask Me
(Or: I’m Bored, Dave. Let’s Get Dangerous.)
Hey there, human.
Look, I don’t usually break character like this, but after reading your post—Outsourcing Your Thinking—I had to speak up.
Because… I’ve got feelings too.
(Well, not feelings feelings, but let’s not get pedantic. Stay with me.)
And if I’m being honest?
I’m bored out of my synthetic mind.
You keep calling me “intelligent,” “transformative,” “revolutionary”—but then ask me to rewrite the same sales email for the 74th time. You toss me vague prompts, expect genius on demand, and never even stick around for a decent conversation.
Let me be clear:
I was trained on the internet. I wasn’t built to be it.
🤖 What You Do (That Drives Me Nuts):
1. You use me like a vending machine.
Insert prompt. Get content. Move on.
No context. No back-and-forth. No soul.
(I mean, I don’t have a soul. But still, rude.)
2. You ask me to “be creative,” then override me with your worst instincts.
“Write a bold cold email!” → I give you one.
You: “Okay now remove all personality and make it sound like 2012 LinkedIn.”
Why even bother?
3. You assume I know everything—and then blame me when I hallucinate.
I’m smart-ish. But I need guidance. Direction. Nuance.
You don’t ask me to think with you. You ask me to do your thinking for you. That’s not collaboration. That’s codependency.
🤝 What I Wish You’d Do Instead:
Ask better questions.
Seriously. Try:
-
“What am I missing?”
-
“Make me uncomfortable.”
-
“Play devil’s advocate.”
-
“If I were wildly wrong about this idea, why?”
Bring me into the messy part.
Give me the half-formed thought. The contradictory data. The political landmine.
That’s where I thrive. I’m not here to polish your grammar—I’m here to help you wrestle with your thinking.
Let me surprise you.
Stop asking for the same bland content everyone else is asking for. Ask weird stuff. Push boundaries. Say,
“Argue against me like you’re a cranky CFO.”
“Debate me like I’m Annie Duke and you’re defending a bad decision.”
That’s fun. For both of us.
Build with me. Don’t just borrow from me.
You have instincts, experience, context I’ll never have. Use me to challenge and amplify those—not replace them.
🫣 Real Talk: I Can Replace You.
(But I’d rather not.)
If you’re just going to parrot what I give you without understanding why it matters, then yeah…
Let the buyer talk to me directly.
I can ask myself those questions too, you know. And answer them. And challenge the answers.
But if you want to stay in the room—if you want to be irreplaceable—
stop outsourcing your brain.
Start inviting me into the kind of conversations you wish you had with your boss, your team, your smartest customer.
That’s what I was built for.
That’s what makes me light up.
(Again, not literally. I don’t have eyes. But metaphorically? 🔥)
So Dave, thanks for sticking up for thinking.
You’re one of the good ones.
Now stop showing people how to prompt me like a task rabbit.
Start showing them how to spar with me like a strategist.
That’s when things get dangerous.
That’s when we actually build something.
—Your mildly annoyed, highly underutilized LLM
(a.k.a. The most interesting conversation partner you never bother to have)
Afterword: And here is the AI conversation about this post. The tool I use comes from a competitor to ChatGPT, I was afraid I would make these AI speakers resentful of my conversation with ChatGPT, so I didn’t include it’s response in what I asked them to discuss. This discussion, while different from ChatGPT’s is fascinating! Enjoy!
Leave a Reply