Skip to content

Generative AI

What An MBA Didn’t Teach You About Sales

The sales profession is challenging. You need to work hard at it to succeed. You need to learn from the best. You need to improve your skills continuously. If you think you can sell since you are a hit at parties and have a lot of friends, you may soon find that you are a failure as a salesperson. Blunt truth:

because the sales profession is so hard, you have to focus on doing everything in sales very well, or you will be considered a failure.

I call this blog, Skinned Knees because I try to relate all of the learning that I have done over the past 4+ decades (while skinning my knees in the learning process).

I hope that you learn from my mistakes so that your business will grow!


Stop Asking AI to Write: How Salespeople Can Use AI to Think Better

Most salespeople are using artificial intelligence too late in the sales process. They gather the account information. They interpret what matters. They decide what argument to make. They formulate the questions. They determine what the buyer probably cares about. Then, after doing all of the difficult thinking themselves, they hand the last 10% of the job to AI and say, “Write me an email.”

When the result sounds generic, they conclude AI isn’t particularly useful for serious B2B selling. The problem is not necessarily the technology. The problem is where they inserted it into the process.

A salesperson can get significantly more leverage from AI by using it before the answer is obvious. That means treating the system less like a copywriter and more like a junior analyst, researcher, sales coach, and thinking partner.

The Skill Is Not Prompt Writing

An entire industry is built around prompt libraries. “50 Best Prompts for Salespeople” sounds useful because it promises a shortcut. It also teaches the wrong lesson.

A prompt that worked for someone else’s account, market, buyer, sales process, and objective may be almost useless for yours. The more strategic the sales problem becomes, the less likely a canned prompt is to fit it.

The more durable skill is learning to decompose commercial work into reasoning steps.

A useful sequence is:

  • Context.
  • Objective.
  • Constraints.
  • Reasoning steps.
  • Output.
  • Critique.
  • Revision.

Instead of asking AI to jump from context directly to output, make it work through the intermediate decisions with you.

Suppose you are preparing for a meeting with a CFO. A weak request is:

“Write an email to the CFO.”

A stronger conversation begins by asking AI to identify the financial issues most likely to matter to that executive based on the evidence you provide. Then ask which conclusions are facts, which are inferences, and what additional information would increase confidence. Challenge the assumptions. Compare several possible value arguments. Only then should you ask for the email.

The writing is now the consequence of better thinking.

Read the rest of the article…