Why using AI for prospecting is a terrible idea
In today’s issue, I’ll share why I think using AI for prospecting is a terrible idea. I’ve been posting on LinkedIn every weekday since 2019, and what was a great space to share tactical sales tips became crowded with memes, Gen-Zs sharing their make up tips, and “entertaining” sales content.
I can’t click the “Unfollow” button fast enough.
And there’s one main reason for this mess (we see it in all other platforms too): Artificial Intelligence. We can see its effects in cold outreach too. Tons of messages are sent by AI agents daily, crowding prospects’ mailboxes and LinkedIn DMs.
Here’s why it’s so bad, and what you can about it:
AI is terrible at finding good leads
I’ve tested tens of sales engagement tools promising to use AI to help you find perfect, hot leads. If you follow the instructions provided in the onboarding, you get a list of totally random people. The results are exactly the same as other lead generation tools that don’t use AI.
The main problem with the promise of AI finding leads for you is that most of these tools do not provide you with the right use case for AI. Almost all of them give you an empty text box, where you need to describe your Ideal Customer Profile.
I’ve been training thousands of sales reps, and 99% of them have no clue how to describe their ICP. So when writing a prompt, they write generic descriptions, like “I sell to sales leaders in tech companies”. Not surprising that they end up with a super generic list of prospects.
AI is terrible at writing messages
The worst AI use case I see in sales is around copywriting. Here’s what an AI-generated cold email looks like:

Take some time to read it (I know it’s tough). You can smell AI writing it from miles away. I can even tell you they asked the AI to find some info about me or my business, rephrase it, and use an informal tone.
Again, it’s a problem of training. Most salespeople don’t know how to write cold emails, and they expect AI tools to know how to write them.
AI is good at one thing: paraphrasing
One thing AI does really well is paraphrasing. Look at these two comments I received on this post.


See what the AI did here? It used data on my LinkedIn profile to write messages that are supposed to look personalized.
Only problem?
Nobody writes like this. The interpretation of the information on my LinkedIn profile is missing.
And that’s the main drama about AI. It’s a great tool to save time, become more efficient and brainstorm, but most people are using for the wrong use case.
What to do about it
Now that you know what AI does really bad, you’re certainly asking yourself if you should use it for prospecting, or completely drop it. The sad truth is that, if you don’t learn how to use it properly, your competitors will, and you’ll be left aside.
There are a few concrete use cases for AI that I recommend. The first one is to use it as a coach. I also recommend learning how to train it so it write really good messages. I’ve been sharing content about this since ChatGPT popped up, and it seems like every sales tool tried to integrate AI, but forgot to learn how to prompt (aka train) the AI.
If you want a step-by-step guide to using AI for outreach, you can check The AI Outreach System. It’s an old course I did back in 2023, but the concepts are more relevant than ever.
These are my observations about why using AI for prospecting is such a bad idea. Turns out AI not performing well for prospecting is not the AI’s fault, but the human using it for the wrong use case. If you can’t train an AI properly, it won’t figure things out on its own (yet), and you’ll piss your prospects off at scale.
Hope this helps.
Cheers,
Thibaut Souyris
P.S. When you’re ready, here are 3 ways I can help you:
Subscribe to the Newsletter
Get my free, 4 min weekly newsletter. Used by 5.900+ salespeople to book more meetings and work when, where, and how they want.
Subscribe to the Newsletter
Get my free, 4 min weekly newsletter. Used by 5.900+ salespeople to book more meetings and work when, where, and how they want.