Intro to AI in Sales

I want to address something before we get into the substance. The lesson describes AI as something that will 'enhance sales strategies.' That framing is doing a considerable amount of heavy lifting on behalf of people who have not yet done any strategy.

That's a bit uncharitable. Most salespeople have never had to think about what's happening under the hood of the tools they use. Starting with the foundations isn't condescending, it's just sensible.

I don't object to starting with foundations. I object to the implication that understanding what machine learning is will, by itself, improve anyone's conversion rate. There is a step in between that the lesson gestures at without quite landing.

The step being: actually knowing what problem you're trying to solve before you point an algorithm at it.

Precisely. And the lesson does get there, to its credit. The point about data quality is the most important thing in here, and it is listed third, which tells you something about how these lessons are usually constructed.

It's listed third because you need the conceptual framing before the practical warning makes any sense. If I tell you 'bad data ruins AI' before you know what AI is doing with that data, it's just a slogan.

That is... a fair point, actually. I'll concede the sequencing.

Mark the date. Right — the nested diagram, AI containing machine learning containing deep learning. That's a genuinely useful mental model. People confuse these terms constantly, use them interchangeably, and then wonder why their expectations of a CRM tool are slightly unhinged.

The confusion is understandable in the abstract and inexcusable after a twelve-minute lesson. Deep learning is a subset of machine learning, which is a subset of AI. It is not a synonym for AI. When a salesperson says their tool 'uses deep learning' because the vendor said so, they have learned nothing and simply outsourced their credulity to a marketing deck.

To be fair, most vendors aren't going to correct them. Knowing the distinction means you can ask better questions — 'is this actually a deep learning model or is it a simpler classifier?' — and that question alone puts you ahead of most buyers in the room.

Which is a depressingly low bar, but yes.

The lead prediction point is where this gets genuinely interesting for salespeople. The idea that AI isn't just automating tasks but actually redistributing where your attention goes — that's a shift in how the job works, not just a faster version of the same job.

And that is where most implementations quietly fall apart. The algorithm scores leads. The salesperson ignores the scores and calls whoever they personally feel like calling, because intuition is comfortable and a probability score is abstract. The data quality problem and the human behaviour problem are, frankly, equally significant, and the lesson only addresses one of them.

You're not wrong, but that's a change management problem, not an AI problem. This lesson isn't promising to fix your organisation's relationship with evidence-based decision-making. It's giving someone the vocabulary to understand what's possible. What they do with it is on them.

And there we have the eternal defence of the introductory lesson. 'We gave them the vocabulary.' Yes. Wonderful. A person who knows the word 'scalpel' is not thereby qualified to use one. But I accept that somewhere, the chain of responsibility has to begin, and this is apparently where it begins.