I'm getting sick of AI-powered outbound


I nearly ended up in a business development meeting with someone who knew virtually nothing about me. His bot, on the other hand, was apparently doing a splendid job.
A rather amusing thing happened recently that sums up much of what is starting to irritate me about AI-powered sales outreach.
I was contacted by someone who appeared to be suggesting an opportunity for joint business development.
That sounded interesting. I'm always happy to explore opportunities where there might be something useful we can do together.
The conversation was convincing enough that I agreed to a meeting.
But something wasn't quite adding up.
I started digging a little deeper and eventually discovered what he was actually offering.
Data centre space.
Now, there's absolutely nothing wrong with selling data centre space. I'm sure there's a healthy market for it.
I'm just not in that market.
Nor could I work out why he thought I might be.
By this point I was getting suspicious, so I asked him a fairly direct question.
"Am I actually having a conversation with you, or am I talking to your bot?"
It turned out his bot had been handling the outreach. He admitted that he knew virtually nothing about me or the businesses I'm involved with.
I had to laugh.
Here was a piece of AI sales automation that had apparently done its job brilliantly. It had caught my attention, started a conversation and persuaded me to agree to a meeting.
There was just one small problem.
Nobody had established whether there was a sensible reason for us to meet.
When a booked meeting becomes the wrong measure of success
I've spent much of my career in technology sales. I know how difficult prospecting can be, and I have no objection whatsoever to using technology to make it more effective.
But think about how my experience might have appeared on a sales dashboard.
Prospect contacted. Positive response. Meeting agreed.
All looking rather encouraging.
Except the opportunity was almost certainly nonexistent.
Had I not asked a few more questions, we might have spent an hour discovering that there was nothing useful to discuss.
I wonder how often that's happening elsewhere.
I don't know, and I'm not going to pretend that one experience establishes a wider industry trend. But it certainly raises questions about what we're measuring.
In my case, the communication worked well enough to persuade me to agree to a meeting. The qualification hadn't worked at all.
And that's quite a distinction.
Personalised doesn't mean understood
My LinkedIn inbox and email are increasingly full of messages that look as though someone has done their homework.
They mention companies I'm involved with, previous roles, things I've written or subjects I've discussed.
Some are remarkably convincing.
But quite often, once I get past the opening, it becomes obvious that the sender has little understanding of what I actually do.
They've collected information about me without establishing whether it's relevant to what they're selling.
AI is making that increasingly easy to do at scale.
It's also making the resulting messages much harder to distinguish from genuine human interest.
The data centre experience was particularly amusing because the supposed opportunity for joint business development was precisely what interested me.
It was only after I questioned the proposition that I discovered the real purpose of the conversation.
The personalisation had worked. The understanding was missing.
Are we automating the wrong part of selling?
I've seen plenty of demonstrations of Claude and other AI tools being used to automate prospecting.
Some of the capabilities are genuinely impressive.
Researching organisations, identifying potential contacts, examining business developments and preparing relevant communications are all things AI can help salespeople do much more efficiently.
I'd encourage salespeople to take advantage of them.
But I'd also encourage them to spend rather more time thinking about the decisions that come before the first message.
Why this company? Why this person? What do we know that suggests there might be a genuine opportunity?
And what would make us decide not to approach them?
Those are basic sales questions. They've been around for decades.
I remember plenty of poorly targeted prospecting long before anyone had heard of generative AI.
The difference now is the speed and scale at which we can do it, coupled with messages that sound increasingly thoughtful even when very little thought has gone into the targeting.
We're becoming exceptionally good at automating the approach.
I'm less convinced we're improving the judgement behind it.
I'd like an AI sales tool that knows when to stop
I'd actually be quite impressed by an AI sales system that occasionally told its user:
"I've researched this person and their business. I don't think there's a credible reason to approach them."
Or perhaps:
"I haven't found enough evidence of a potential fit. You might want to investigate further before contacting them."
I'd regard that as valuable sales intelligence.
It wouldn't increase the number of messages sent or meetings arranged. In fact, it might reduce them considerably.
But it could improve the quality of the conversations that actually take place.
Of course, selling isn't an exact science. Some of the best opportunities emerge from conversations where neither party initially knows what might develop.
I certainly wouldn't want technology to eliminate that possibility.
But there is a difference between exploring a plausible opportunity and sending someone a convincing message when you haven't established any meaningful reason for contacting them.
AI ought to help us make that judgement.
So what are we actually celebrating?
I keep thinking about the chap selling data centre space.
I'm sure he wasn't deliberately trying to waste my time. He had a proposition, he'd invested in automation and his bot was getting results.
At least by one measure.
He got someone to agree to a meeting without knowing much about them.
That might be a success for the automation provider. It might even look like a success for his sales team.
But was it good selling?
I don't think so.
As AI becomes more deeply embedded in sales, I'd like us to look beyond the number of messages sent, responses generated and meetings agreed.
How many of those conversations were worth having?
How many created a genuine opportunity for either party?
And how many should never have happened?
I'm all for using AI to make salespeople more productive.
I'd just prefer we became better at deciding who to sell to before becoming extraordinarily efficient at contacting everyone else.




Comments