What’s All The Fuss, Prompt Engineering Isn’t New…..

My feeds are filled with the miracles and mysteries of prompt engineering. I’m attending an AI for Sales Conference, where the underlying theme is “we have to master prompt engineering.”

I scratch my head, thinking, “What’s the fuss? I’ve been doing prompt engineering all my career. We just never called it that!”

Not long ago, I was contacted by the RevOps/Enablement team of a very large SaaS company. They had been exposed to some of the AI tools we provide many of our clients. Initially, they said, “You are using AI in a very different way than most. We’d love to learn what you are doing and the prompting tools you use.”

In one of the initial meetings, someone asked, “How did someone like you get into prompting at such a sophisticated level?” At first, I wondered whether I should take offense, were the referring to my white hair and old fart status? Possibly they were wondering, “You’ve held executive roles in large companies and been a strategic consultant for so long. No one with that experience could ever have a clue about prompting!”

They were stunned when I gave them that line, “I’ve been doing prompt engineering my entire career, we just never called it prompt engineering.”

They asked me to explain, suggesting I go through key experiences/examples in my career where “prompt engineering” stood out.

I went way back–when I was studying for a PhD in Applied Physics. I was deeply involved in research, new applications of the Heisenberg Uncertainty principle. Scientific research is governed by one thing above all: the Scientific Method.

• Clarity of intent/objectives: What are the goals of the research, what are we setting out to prove, disprove, discover?
• Context/background setting: We had to provide the framing or background of why we were conducting the experiments, where the results might be positioned within the overall framework of scientific knowledge and research.
• Constraints and guidance: Every experiment had clear constraints. There were boundaries, things that we would have to exclude in our experiments. There may have been certain methodologies or processes we had to follow.
• Step by step reasoning: This is related to the constraints and guidance, but we had to provide clear step by step reasoning, both to guide our own work, but more importantly, for others to validate and verify our work.
• Iterative refinement: Inevitably, as we conducted the research and experiments, we would refine our approaches, even refining some of our original objectives.
• Tone and framing: In research, how we framed our hypotheses and positioned our results had a huge impact. Framed incorrectly, we would go down the wrong path. The tone and framing of our results, had a huge impact on how they were received by people reviewing and applying our work.

I’ll stop here, there is much more, but this gives you an idea of what any scientist does in conducting research and experiments. The specifics to each item vary, but the underlying framework and processes are always based on these fundamentals.

I made a career switch (that’s a whole different story). I went into selling–selling mainframe computers for IBM. As I sat through weeks of training, my nerdy mind started thinking, “A lot of this is very similar to what I did as a research scientist. It’s just applied in a very different way.”

We learned things like: The sales process, methodologies, deal execution, call planning, account/territory planning, pipeline management. And then specific skills like qualifying, discovery, presenting, objection handling, closing and others.

As I looked at everything we learned I could see the same underlying foundational principles used in scientific research. For example, just in planning a high impact sales call, we went through the same process:

• Clarity of intent/objectives: What are our and the customers’ objectives for this meeting? Are we aligned with what we are trying to achieve.
• Context/background: What has happened prior to this particular meeting? What got us to this point? What has created the need for this meeting.
• Constraints: Who is participating? How much time do we have? Are there limitations to what we might be able to explore or may be out of scope for the meeting?
• Step by step reasoning: Using the selling and buying process as a background, how does this meeting fit in? How do we use it to move forward with the customer.
• Iterative refinement: What have we learned in past meetings that might cause us to adjust how we conduct this meeting?
• Tone and framing: How do we connect effectively and impactfully with the people participating in the meeting. Should we frame certain elements we discuss as being critical? What tone do we establish in communicating with the people in the meeting.

As we look at everything we learn and do in selling the foundational principles outlined above provide the basis around which we develop our account, territory, prospecting, sales/buying process, ICPs, deal strategies, call plans, retention/upsell plans.

You may be starting to get the idea, prompt engineering is nothing new or unique to AI. The fundamentals of great prompt engineering is something we have all learned and done before. We just never called it prompt engineering. It may have been the scientific method, our sales process, our methodologies. We may have labeled them Challenger, MEDDICC, Solution Selling, Winning By Design, GAP Selling. But each of them are built with the same foundational elements.

But let me get back to my career journey. I moved into senior and executive management roles, I dealt with engineering, manufacturing, ops, finance, and other organizations. Each of these were very different, but each of them had foundational frameworks around which they built their strategies and executed the work. I won’t take you through each of these, I hope you are getting the idea.

Even in our consulting practice, we have certain processes, frameworks and methodologies we use in each project. The clients and projects are very different, but it’s the understanding and consistent execution of these things that enables our clients to achieve profound results.

Prompt engineering should not be a mystery. Each of us has been doing it, perhaps unconsciously, through our careers. Our ability to do these things has enabled us and our organizations to achieve profound results.

Perhaps one of the most important things we can learn in our AI prompt engineering is not what the gurus would have you learn about the “mysteries of prompt engineering,” Rather it’s rediscovering the fundamentals we’ve been practicing for years, and applying those in how we “prompt” and leverage AI.

Afterword:  This is an outstanding AI based discussion of this post.  I’ve noticed the discussions seem much better when they are talking about AI.  I suppose that’s natural, we love talking about ourselves  😉  Enjoy!