How to use AI for B2B prospection
AI is not going to replace salespeople, but it might replace those that don't use AI
3/25/20263 min read
AI in B2B Prospecting: Is It Actually Useful?
Today, you'd have to be living under a rock not to know what's going on with AI.
On the one hand, you have all these startups claiming they can replace salespeople with AI agents. On the other, there's all the hype surrounding the latest Claude models and new AI tools (OpenClaw, etc.). And then, somewhere in between, there are a few things that actually deliver real value.
Here's a spoiler: an AI agent that supposedly prospects 24/7 with incredible efficiency will most likely be incredibly efficient at spamming your prospects with inaccurate, irrelevant, or poorly personalized messages, potentially even sending them to the wrong people.
As for the latest Claude models and other new AI tools, they mostly give LinkedIn and TikTok influencers an opportunity to rack up views, without necessarily offering any real added value beyond the hype (at least in most standard use cases).
However, where AI can actually make a difference in B2B prospecting today is when an expert builds custom workflows using the latest LLM models for a specific market, or when traditional growth marketing platforms (such as Clay, Apollo.io, etc.) integrate AI to make their tools more powerful and improve their databases.
In this article, I'll explain how I've used AI to automate my B2B cold email campaigns with pretty good results, as well as the different ways you can use AI to save time and improve your prospecting performance.
This AI-enhanced approach to email and LinkedIn sequences is something I can also look to replicate for my clients as part of my Outbound Machine offer.
Custom AI Workflows
Today, one of the most valuable applications of AI in B2B prospecting is account research and writing personalized email and LinkedIn messages.
However, getting good results requires solid expertise in your field and a deep understanding of your market, so you can properly guide the AI with precise prompts.
It also requires some growth marketing knowledge to build your own workflows (I personally use n8n). Because, let's be honest, fully automated, plug-and-play AI agents that supposedly do everything for every industry are simply not good enough yet.
a) Account Research and Competitive Intelligence
If you have a good understanding of your addressable market, you've probably already segmented your prospects into several lists: high-priority targets, interesting prospects, and lower-priority accounts.
To approach high-priority prospects as effectively as possible and increase your chances of success, you can use AI to research individual accounts: recent news, potential angles based on different buyer personas, and so on.
You can also ask AI to identify what differentiates you from your competitors and what aspects of your offer might appeal to different buyer personas.
b) Writing Personalized Messages
The more personalized a B2B cold email is, the more relevant it feels to the recipient, and the more likely it is to generate a positive response or book a meeting.
However, writing highly personalized emails takes a lot of time, especially when you need to send at least 100 emails to generate one meeting (a 1% conversion rate is already considered very good in cold email outreach).
AI can write highly relevant messages, provided you give it the right data and a well-designed prompt.
But finding the right data means knowing exactly what information you need to collect to make your message more relevant in the first place.
That requires understanding how your prospects think, what makes them react, their recurring objections, their misconceptions, and so on.
In other words, if you know exactly which variables, insights, or metrics are most likely to resonate with a prospect, you can use them to create highly personalized messages.
And when it comes to understanding your target audience, there's nothing quite like cold calling. It's the only prospecting channel that allows you to truly understand your prospects, because you're actually interacting with them directly.
Ironically, cold calling can therefore help you create much better email and LinkedIn campaigns.
c) A Real-World Email Example
All of this might sound a bit theoretical, so let me give you a concrete example.
In my industry and target market, I know that the things most likely to get prospects interested are:
The number of meetings I could potentially generate in their market.
The companies I've previously worked with in their industry.
Whether I can build relevant prospect lists for their business.
So, if I want to send a personalized message with the highest possible chances of generating a meeting, I can address these points directly.
But, of course, this requires having access to several pieces of information beforehand:
Who are their target customers? Large enterprises or small businesses? Marketing Directors or Procurement Managers?
Which relevant case studies and social proof can I use? Companies I've worked with in their industry, or businesses targeting a similar audience?
I collect this information in advance using custom enrichment workflows I've built with n8n, Apify, and Icypeas.
I then feed the results into an AI model responsible for writing the messages according to a specific format and set of instructions I've defined.
This allows me to generate a significant number of meetings, potentially at scale (you can send a large volume of emails per day if you want to), while keeping costs relatively low.