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

We transformed our standard dev email flow into an AI-powered conversion powerhouse. 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 integrate AI into our workflows, we knew we wanted to achieve 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 needed to determine which AI applications would have the greatest impact.

Email marketing seemed like a natural place to start. We had been running optimization tests on our development workflows for years, but after a while the gains became incremental by a few percentage points here and there.

We needed something that would be a total game changer. Something that has had both a significant influence on top-of-funnel metrics and practical usability within 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 AI to transform our email performance. We will also explain how we achieved an 82% increase in conversion rates – and all the lessons we learned along the way.

Identify the challenge

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, then tailoring content to those groups.

For example, if someone uploaded marketing-related content, we would send them more marketing assets rather than sales content.

It wasn’t a bad approach. But it was mostly educated guessing at the group level. We said, “People like you usually want this,” rather than understanding what each person was trying to accomplish.. We wanted to do better than that.

The hypothesis: moving from groups to individuals

The more we looked at AI…