Inbound

December 22, 2025

How to Use AI to Personalize Emails at Scale

If there’s one skill marketers should commit to mastering in 2026, it’s using AI to scale 1:1 email personalization.

At INBOUND, Ari Echt-Wilson, Director of Global Prospect Email at HubSpot, shared how her team has driven over 10,000 qualified meetings per quarter and achieved a 45% year-over-year increase in email conversion rates with the help of AI-personalization.

Here’s what she and her team learned throughout the process and the actionable skills every marketer can start practicing now.

HOW HUBSPOT AUTOMATES 1:1 EMAIL PERSONALIZATION WITH AI

Echt-Wilson and her team have built an automated workflow that uses AI to generate truly personalized emails at scale without requiring reps to write each message manually. Here's how it works:

When a contact is added to a workflow (like after visiting a pricing page) their website gets scraped to learn what the company does.
It then pulls in CRM context like job title, conversion history, page visits.
That full profile gets sent to an LLM (like OpenAI), along with a prompt asking for a relevant, tailored email.
The model returns a unique message, which is stored in the contact record and inserted dynamically as a personalization token in the final email.

Now, let’s break down the skills, systems, and strategies her team uses to make it all work.

PERSONALIZE EMAILS WITH PURPOSE

Not all personalization is good personalization. Real personalization is about relevance and AI can help you get there faster, if you feed it the right context.

Echt-Wilson showed how her team pulls from CRM data, page views, and even chat transcripts to give AI the full story before it ever drafts a line of copy.

Precision-level personalization increases engagement, improves click-through rates, and drives higher-quality conversions especially in industries where timing and context matter.

Try this: Before prompting email copy, ask yourself: what would a great rep say on a call? Use those inputs as your prompt foundation.

COACH AI LIKE A TEAMMATE

AI needs direction. It can write emails but without thoughtful coaching, the outputs are often long, vague, and not helpful.

“Like any great athlete, AI needs a great coach.” — Ari Echt-Wilson

Her team breaks each message into modular sections (like opening, pain point, solution), writes detailed prompt instructions, and layers in clear tone guidelines. This also means building workflows that make it easy for marketers to evaluate outputs, update prompts, and keep improving over time.

Try this: Write prompts like creative briefs. Be clear about structure (Pain > Solution > Result), tone, word count, and voice. Then, share example outputs that performed really well with the AI model you’re using.

ITERATION IS THE STRATEGY

The best-performing emails you’ll write with AI probably won’t be the first ones.

“You should know that we have failed at every single new AI use case on the first try.” — Ari Echt-Wilson

Echt-Wilson’s team tests, refines, re-prompts, and tests again. Every iteration is a feedback loop. They typically test outputs on 100+ sample contacts before sending anything live, reviewing and refining prompts until the results are consistent.

“It takes three plus iterations on average to see results. And the magic happens after a lot of teams give up,” Echt-Wilson shared.

Stick with it. AI email performance improves with reps. The more you use it, review it, and retrain it, the more reliable it becomes.

Try this: Treat every email you send as training data. Feed back top performers and tweak your prompt based on what actually converted.

CLEAN DATA IS THE DIFFERENCE

Bad data leads to bad emails. One of Echt-Wilson’s examples was an email congratulating someone for being a COO, who turned out to be an intern. Oof.

Clean, updated CRM data makes personalization powerful instead of awkward. It also opens up more opportunities for nuanced segmentation and dynamic content. Once her team added unstructured data like chat transcripts and customer activity, they saw a 76% increase in conversion rate. For any team investing in AI automation, better inputs lead to much better outcomes.

Try this: Prioritize context (like recent activity, industry signals, or account size) over quantity when it comes to customer data.

BLENDING ART AND SCIENCE

Echt-Wilson wrapped her session with an artful analogy. When photography was invented, people thought it would kill painting. But it actually sparked new artistic and creative movements. It forced artists to focus more on emotion, style, and interpretation than replication.

We’re at a similar moment. Just as artists evolved, great marketers will find their new creative stride by leaning in and evolving quickly with AI.

Watch Echt-Wilson’s full session below: