How to use AI to 4x cold email reply rates

In B2B sales, we have to know our customer. How can AI agents help with personalization?

💪 Welcome to “Nailed It”, a series where we interview real people who use AI to improve everyday work and life. 🤖 With many AI tools out there, we’re seeing lots of hammers in search of nails. ✨ We're on a mission to uncover the AI power users who nailed it.

---

If you do any type of cold email outreach (sales, job search, etc), you know that personalization is key. Your prospect is more likely to reply if the message feels tailored to them and speak to their pain points.

However, personalization is easier said than done. True personalization takes lots of research, which can be very time-consuming—Sounds like a great job for AI!

This week we talked to Michel Lieben, the CEO of ColdIQ (a sales prospecting agency). He is the master of using AI tools to automate outbound sales.

Let's say he's selling to e-commerce companies, and he wants to personalize each outreach email to mention a product the company sells.

Here's his current AI workflow:

1. Input the list of companies in Clay, a data enrichment tool powered by AI. Use Clay's "Claygent" web scraper. Write a prompt using natural language, instructing the web scraper to find a product each company sells.

Clay will then automatically fill the sheet with a product name of each company:

You can use the same feature to find a strength of each company you're targeting:

2. Use the cold email automation tool Smartlead to personalize the outreach emails with the data from Clay (Clay has an integration with Smartlead).

For example, include a line that speaks to the company's strength. This line would look different for every prospect, thanks to the data enrichment we did earlier.

Using this AI-empowered personalization approach, Michel has been able to improve his email reply rates by 3-4 times—not to mention the massive amount of time saved on manually scraping data online.

📝 Lessons learned

• AI can save you hours on account research, but it's still up to the human to decide how to approach the personalization. Bad personalization is generic and irrelevant; good personalization is specific to the product you're selling and has an impact on the rest of your pitch. This decision is something that can't be automated with AI.

• As Michel suggests, "Don't use AI for the sake of using AI". The correct approach to integrating AI into your workflow is: Map out your process, find out which steps could be more efficient, and then look for relevant AI tools. If you can find an AI solution that will help you, then amazing. If it's not helpful, don't force it.

Watch the full YouTube video here!