AI in Sales and Sales Training

I've been thinking a lot about AI in sales recently. Not so much about what it can do (we know it can do loads) but about what needs to be in place for us to use it well.

AI can research accounts, summarise information, prepare for meetings, analyse opportunities, develop questions, challenge assumptions, draft proposals and review conversations.

And yes, it can write emails.

But I think we're in danger of starting the conversation in the wrong place. Because before we ask what AI can do for a salesperson, there's another question worth asking:

Does the salesperson know what good looks like in the first place?

AI doesn't remove the need to think

I've always believed good selling is largely about good thinking. Thinking about the right customers, the problems they might be trying to solve, who is involved and what matters to them. Thinking about the questions worth asking, value from the customer's perspective and what needs to happen next.

Sales techniques, models and tools give us ways to structure that thinking. AI now gives us an extraordinary opportunity to enhance it.

But there's an important order to this.

You need to know how to think before you can enhance your thinking with AI.

Otherwise, you're simply outsourcing something you didn't understand in the first place. And that's potentially dangerous because AI is remarkably good at producing something that looks convincing.

A mediocre salesperson with an LLM can produce a beautifully structured mediocre proposal in seconds. They can write a polished prospecting email to entirely the wrong person, generate 20 discovery questions without understanding why any of them are worth asking, or create an impressive account plan based on assumptions nobody has bothered to test.

Faster isn't necessarily better. Sometimes it's just faster.

As the saying goes: A fool with a tool is still a fool.

AI isn't another sales module

This is where I think some conversations about AI in sales training are going wrong.

There's a temptation to treat AI as another module. We teach prospecting, discovery, presentations and negotiation, then somewhere towards the end we add: “And now here's how you can use Copilot.”

I'm not convinced that's the answer.

AI isn't another sales skill. It's increasingly something that can support the application of all those skills. But the salesperson still needs to understand what they're trying to do.

Take account research. AI can save an extraordinary amount of time researching hundreds of potential accounts and help identify the relatively small number that appear to fit your Ideal Customer Profile.

Brilliant... but only if somebody has done the thinking required to define a meaningful ICP. Otherwise we've simply automated vague targeting.

The same applies to discovery. AI can help me prepare for a meeting, explore possible customer problems, identify stakeholders and develop potentially useful questions. But I still need to understand what good discovery looks like.

I need to listen, notice when something doesn't quite add up, know when to probe and when to shut up. I need to recognise the difference between something that's merely interesting and something that's commercially important.

AI can support that thinking.

It doesn't absolve me from doing it.

Do we really need all those tools?

There's another thing I've noticed. We seem to be accumulating an awful lot of AI sales technology.

Every week there appears to be another spangly new tool promising to transform prospecting, automate research, coach conversations, generate content or revolutionise selling in some other previously unimaginable way.

Some of them are excellent. But I do occasionally wonder how many we actually need.

A good LLM can already do an enormous amount if you know what you're trying to achieve, what good looks like, what information matters, what context to give it and how to judge what comes back.

That last bit matters enormously.

The better AI becomes at sounding authoritative, the more important human judgement becomes. If you don't know enough to challenge the answer, a confident answer can very quickly become a confident mistake.

I'm becoming less interested in prompts and more interested in context

Imagine asking an LLM:

“Write me an email to an engineering director.”

It will oblige. It'll probably be perfectly well written. It'll also probably sound remarkably like the other 700 AI-generated emails that engineering director received this month.

Now imagine the AI already understands your ideal customer, your proposition, the problems you solve, the language customers use, the outcomes you've created, the stakeholders typically involved, the evidence you have, the reasons customers change and the reasons they don't.

Then add what you've already learned about this particular account.

That's a completely different starting point.

Which is why I'm becoming less obsessed with clever prompts and much more interested in good context.

The clever bit isn't necessarily the prompt.

The clever bit is the thinking that created the context.

From workbook to Sales Brain

This is also why I've become increasingly interested in building what I call 'Sales Brains.'

For years, sales training has tended to produce a workbook. People complete exercises during the programme, record some thoughts, perhaps refer back to it afterwards and, if we're being realistic, quite often put it somewhere and never look at it again.

There's nothing inherently wrong with a workbook. But AI gives us the opportunity to think differently about what those exercises are actually for.

What if the exercise isn't simply there to demonstrate learning? What if it's capturing useful commercial intelligence?

Your ICP. Your proposition. Customer problems. Change drivers. Stakeholders. Value. Messaging. Evidence. Questions. Outcomes.

Now those separate pieces of thinking can become context for AI.

The training helps people develop the thinking. The exercises structure and capture it. The Sales Brain gives AI the context. And the salesperson can then use that context when they're preparing for an actual meeting, researching an actual account, thinking through an actual opportunity or working on an actual customer problem.

For me, that's a much more interesting application of AI in sales training than simply showing somebody how to get Copilot to write an email.

The workbook records what you learned. The Sales Brain remembers what matters.

Perhaps AI makes sales capability more important, not less

I'm not remotely suggesting we should throw traditional sales capability away because AI has arrived.

Quite the opposite.

I think AI makes genuine sales capability more important. People still need to understand customers, be commercially curious, discover problems rather than leap to solutions, understand value, conduct collaborative conversations and think commercially.

Perhaps most importantly, they need to recognise when what they're being told (by a customer or by AI) deserves another question.

Then AI can help them do more with that capability. It can research faster, explore wider, prepare better, remember more, spot patterns, challenge thinking and reduce administration.

And hopefully that gives people more time for the bit of selling that still matters enormously:

Having great conversations with customers.

That's the bit I don't want us to lose.

AI shouldn't be about removing thinking from selling. It should be about giving good salespeople greater capacity to think and helping them apply that thinking more consistently.

So before investing in the next spangly AI sales tool, perhaps there's a simpler question worth asking:

"If I gave my salespeople the world's best AI tomorrow, would they know enough about selling to use it well?"

Because if the answer is no, I'm not sure you have an AI problem.

You might have a thinking problem.

And a fool with a tool is still a fool.