Welcome back to the GTM Engineer Experiment Series. If you’re new here, a big part of my job working in growth and GTM Engineering at startups like Hearth, Rippling, and now Netic is trying out a lot of new products and running experiments. I was an early user of many now popular GTM engineering tools like Clay, HeyReach, and PhantomBuster.
Since GTM alpha comes primarily from ideas, software, and workflows that others haven’t found yet, I try new tools and ideas often.
In this series, I (and members of The GTM Engineer Lab) document what works, what doesn’t, and what we think is really going on. You should be able to take learnings from these experiments and apply them to your own job (and you’ll have to tell us what works! Just email us at hi@thegtmengineer.ai).
Today, I’m sharing an experiment from Cory Nelson, Director of Marketing at Harmonya, which is a product intelligence platform for consumer packaged goods (CPG) brands and retailers with $20M+ raised. He uses Brew to run an intricate email (often known as lifecycle) marketing program with just 5% of his time that more than doubled their engagement rates.
For as much as there is too much talk about cold outbound email, there is equivalently too little talk about email (or lifecycle) marketing.
Email/lifecycle marketing is sending email campaigns into an opted-in list of prospects and/or customers. The goal is typically to turn cold buyers into warm ones, warm into hot, hot into customers, and customers into upsells. It can also improve your brand reputation and add real value to your TAM.
Since Harmonya sells into enterprise brands like PepsiCo and Colgate, and they have a material opted-in list, email marketing is an important lever for their team. Staying top of mind and getting the right message to someone at the right time could turn into a massive deal.
Historically, though, email marketing was very high friction, so Cory (the only marketer) would send a monthly newsletter and occasional product announcements. Each time he wanted to send an email, he would go into Hubspot, build the email off of a template, and select Harmonya’s entire mailing list.
Today, Cory orchestrates email campaigns to his mailing list using Brew within Claude. It segments his mailing list by persona, funnel stage, pain point, and sub-industry in seconds. It drafts messaging for each segment, and it creates full multi-step email sequences with A/B testing to improve copy and subject lines over time.
Though Cory is spending no more time on lifecycle marketing than he was before, it’s gone from a channel where he was sending occasional updates, to one being run like it has a full-time marketer in seat.
In this essay, I’ll walk through precisely how Cory set up and built this engine with Brew, and how you can do the same thing at your company.
As I dug into this experiment with Cory, what stood out to me is how under-resourced lifecycle marketing software has been. While there seems to be an AI tool for almost every part of sales, marketing, and GTM data, this category (lifecycle marketing) is surprisingly absent.
Running lifecycle campaigns has historically been pretty miserable: manually filter for your audience, write the message, design the template, do manual CSV list uploads, and manually check back in on results. This has all taken place inside of either a standalone, disconnected tool or an overly expensive, hard-to-use legacy product.
It should behave in a completely different way: inside of an LLM, dead simple to segment, and custom + fast to design on top of your existing brand materials (and web research). This is exactly what Brew has built and why I was excited to partner with them on this experiment. Thanks to Brew for supporting and sponsoring the GTM Engineer community.
So what is Brew?
Lifecycle marketing involves sending emails to people who have already opted in to hear from you. At its worst, it’s sending emails to check a box. At its best, it’s segmenting your mailing list so that everyone gets content relevant to them, and you stay top of mind to drive free demand over time. Most teams run lifecycle by hand out of platforms like Hubspot or Marketo, leaving the channel under-engineered.
Those platforms were built for a person to click through a builder. Every step is a human operating software, and every additional audience segment they want to build, email they want to send, or design template they want to create adds a linear amount of work.
Brew is an AI-native email service provider. You can describe a campaign you want to launch, or content you want to share, and Brew builds all of it: the audience, copy, email design, and multi-step sequence logic. This all sits automatically on top of your existing brand assets. You can either use Brew’s native ESP or push through your existing provider (Mailchimp, HubSpot, Klaviyo, Marketo, etc) to run way faster. Plus with Brew’s MCP, you can run everything out of wherever you work now, whether that be Claude, ChatGPT, or Cursor.
Why Cory decided to use Brew
Cory already used Claude for some of his marketing work like one-pagers and decks. However, when it came to email, he was still doing everything manually in Hubspot.
After finding Brew, he handed over recent content he’d published and asked it to create a campaign. Thirty seconds later, it produced an email sequence Cory liked, and when he asked for three variants with different CTAs and subject lines, he got those back, too.
Next, he pointed it at Harmonya’s white paper on agentic commerce, along with a blog post and video. He told Brew he wanted a short sequence using all three assets along with an A/B test at each step.
Brew returned a thoughtful multi-step sequence. Email one was zero click and anchored its messaging around a few stats, email two drove to the blog post, and email three handed over the full white paper. What’s more, Brew built complex underlying logic with intentional wait periods between sends and branching that only routed the follow-up emails to people who opened the previous ones.
The output wasn’t just an improvement to what Cory was doing, it was completely new. Previously, Cory only had the time to draft one email that got sent to the whole mailing list, so this felt transformative.
Today, Brew sits on top of HubSpot at Harmonya. It cuts their mailing list into segments by variables of Cory’s choosing like persona, industry, and funnel stage. While their mailing list isn’t as large as a B2C business, they have some of the largest CPG brands in the world opted in. Those brands are different enough that the same piece of content and messaging rarely resonates with everyone.
Now with Brew, if Harmonya publishes a blog post about a new trend in fiber, Cory can tell Claude to find who to send it to. Then, using Brew, it will segment the lists based on the industries of the recipients (cereal, salty snacks, canned goods companies, etc) and draft distinct emails about fiber’s impact within each sub-category.
Not only is Cory building campaigns he wouldn’t have dreamed of before, but the new campaigns have more than doubled open and click rates versus what he was previously getting.
Exactly how Cory built this
Once Cory decided to commit to Brew, here’s exactly what he did, in detail:
Set up Harmonya’s brand and the sending domain
Brew’s brand builder crawls your site for everything about your brand like voice, positioning, colors, typography, layouts, imagery, and content. If you have existing templates/designs you’d like to use, you can also import those from Figma, HTML, images, and even email forwarding. This way, you don’t need to go to your design team (or ask Claude for custom HTML) when you want to create custom headers for your email. And, when you want new design templates, instead of getting AI slop, it is rooted on top of your existing style and assets.
After importing the brand details, Cory set up the tracking and CNAME record.
Connected HubSpot and added Brew’s MCP server to Claude
The MCP server exposes Brew’s full toolset. You can draft messages, create audiences, send emails, stand-up automations, and analyze data.
Cory does most of his email building directly in Brew since its UI makes it easy to visualize everything, but he does most of the back-end automation, sending, and data tracking via Brew’s MCP inside of Claude.
example of Brew in action
Described the campaign he wanted instead of building it
He starts by giving Brew a message or piece of content he wants to distribute. He can give as much (or little) guidance as he wants here. Brew has embedded taste based on what works best, but it can also adapt to how he wants to run his lifecycle motion.
Brew then returns emails, email variants for A/B testing, and sequence logic. Cory will sometimes make tweaks here, and sometimes ship it as-is.
He then asks for Brew to build the audience. It does the heavy lifting of creating filters for segments, and ensuring lists don’t overlap. This step used to take a ton of manual effort, but now it’s just one more prompt.
Reviewed every piece of the first campaign to ensure Brew was working
Cory carefully read through every piece of copy in the first campaign looking for AI hallucinations, but found none: “it’s self-contained within the assets that I gave it. It wasn’t making up new stats and things like that.”
Now he spends less time reviewing and still has confidence when shipping new campaigns to his enterprise buyers.
Started analyzing campaign results
When every email was sent to the whole opted-in list, there wasn’t much data to go look at besides open/click rates. Now, Cory has useful data with every campaign’s distinct variants. He can use this to not only improve his future lifecycle emails, but also inform how he thinks about positioning across all of Harmonya more broadly.
Cory reviews results and bounce rates in Claude via Brew’s MCP. Then, he has it improve future campaigns by assessing why certain email variants perform better.
Began running a full lifecycle program with the same 5% of his time he was spending on email before Brew
Cory gives Brew ten or so content links and has it handle formatting and sending for the monthly customer newsletter.
It quickly generates segmented campaigns when the team creates new marketing content.
It also checks which contacts haven’t heard from Harmonya in a while, and he has Brew determine which email they should get next based on their lifecycle stage.
The six steps took Cory from a manual process to get even a few emails out a month, to describing lifecycle campaigns that Brew executes at scale.
Why Brew is uniquely suited to support this experiment
Three features make Brew particularly useful for this experiment:
1). It can do every step of the job: Most AI email features just write you a draft. Brew covers all tasks. It drafts copy, designs emails based on your brand, creates segments, wires together campaign logic, verifies addresses before a send, and ramps volume to warm the domain being used. Because none of those features are missing, Brew can handle all elements of execution, which lets Cory focus his energy on choosing the campaigns he wants to run and messaging angles he wants to hit.
2). An agent can run all of it: Brew’s MCP server exposes its full toolset, and because it can connect and read from your existing stack (HubSpot in Cory’s case), an agent has everything it needs for lifecycle marketing. That’s why one prompt from Cory in Claude can get an agent to segment the audience and do all of the work to draft and launch a campaign. An agent pointed at a tool that’s blind to your own data can’t execute a build end to end on its own.
3). The first AI-native lifecycle email tool: Brew is the first fully AI-native email lifecycle tool I’ve seen. With so many slow and manual tasks required to do lifecycle well, it’s a no-brainer for anybody that’s serious about scaling the impact of this channel. Don’t forget, the ROI is nearly infinite since there’s no inherent acquisition cost!
This experiment worked because the issue with running a lifecycle program was never coming up with strategy. It was justifying the time it would take to produce multi-step campaigns with many segments, run analysis on their results to improve following iterations, and maintain all of the infrastructure underneath.
Final thoughts
Lean marketing teams often have to choose channels to deprioritize. Everyone agrees they’re valuable and should be live, but they sit untouched because whoever would own them is already juggling multiple higher-priority items. Lifecycle emails are often one of those channels, because returns from the emails aren’t immediate, you can only scale them by growing your opted in list, and most execs don’t see the potential in this channel.
What I like about Cory’s build is that he didn’t try to go hire someone to take over his email responsibilities. Instead, he figured out how to get way more done with the same 5% of his time he was already spending on email.
If you want to set up something similar at your company, here’s what matters most from Cory’s experience:
Implement AI-native tools to increase quality & velocity on channels you believe in, but are capacity constrained on
Since Cory was already running email marketing, he knew exactly what he would do if he had more time. When he got Brew, it gave him the execution muscle he needed to scale the impact quickly. If there’s a channel you’re running but know isn’t optimized because you don’t have the time, there’s a good bet that tools like Brew can help you.
Start by distributing an asset you already have
Cory’s first campaign came from a white paper, a blog post, and a video that were already published and paid for. He used the collateral to test if Brew could better package the content to Harmonya’s audience. Once it worked, he knew he could lean on Brew as a distribution channel moving forward.
Segmentation only works if the data upstream is already clean
Brew can write a different subject line for each industry Harmonya sells to because their HubSpot already has enriched data. Cory did that work before implementing Brew. If you want to have an agent segment your list, it needs accurate data in the first place.
Make sure to review the agent’s work
Cory read every email in the first campaign Brew produced. Agents can be a super effective execution tool, but when they’re working on content that will touch customers and buyers, it’s important to verify the copy meets your standard while you are training the AI on your brand voice and positioning.
Don’t set it and forget it.
Cory still spends 5% of his time on lifecycle marketing. He picks which assets plug into lifecycle campaigns, gives Brew links to the content, and reviews the sequence structure and copy before pushing anything live. A lifecycle program with nobody’s judgment + taste produces slop emails at high volume, which will drive unsubscribes and mistrust in your audience.
Thanks for reading! Let me know if you try this out and reply to this email if you have a GTM Engineering Experiment you’d like featured :)
Lifecycle marketing running on five percent of someone's time while doubling engagement is the detail that should get more attention than it does, because it implies the previous version was not underperforming from bad copy, it was underperforving from neglect. Most lifecycle programs die from nobody owning them daily rather than from a bad sequence, so a tool that makes ownership cheap enough for one person to sustain solves a staffing problem disguised as a marketing problem.