How We Used AI to Increase HubSpot Email Conversions by 82%: A Case Study

We’ve turned our standard nurturing email flow into an AI-driven conversion drive. Here’s what we did, what worked (and what didn’t), and what we learned along the way.

When our marketing team started discussing how to strategically incorporate AI into our workflows, we knew we wanted big results. But here’s the thing about big results: they don’t come from trying everything at once.

With limited resources and unlimited possibilities, we had to hone in on which AI applications would have the greatest impact.

Email marketing seemed like a natural starting point. We’ve been running optimization tests on our breeding streams for years, but after a while, the gains have become incremental by a few percentage points here and there.

We needed something that would completely change the game. Something that had a significant impact on funnel start metrics and practical usability in our marketing team. But what — and how?

In a recent Marketing Against the Grain episodeHubSpot VP of Marketing Emmy Jonassen and I share how we experimented with artificial intelligence to change the performance of our emails. We will also explain how we achieved an 82% increase in conversion rate. — plus, all the lessons we’ve learned along the way.

Identifying challenges

First, let me explain what we did before AI. Like most marketing teams, we approached email personalization through segmentation — grouping prospects based on similar characteristics and then tailoring content to those groups.

For example, if someone downloaded marketing content, we would send them more marketing resources instead of sales content.

It wasn’t a bad approach. But it was essentially educated guesswork at the group level. We’ve been saying, “People like you usually want that,” instead of understanding what each individual person is trying to achieve. We wanted to do better than that.

Hypothesis: Transition from groups to individuals

The more we looked at the AIs…