Customer showcase: How a 3-person team uses AI to scale global email marketing without sacrificing quality

How do you manage high-volume email marketing across multiple languages and markets with a team of just three people, without slowing down production or letting mistakes slip through?

For one large enterprise client at Underground Ecom, the answer was rethinking quality assurance with AI.

The results? Over six months, the team:

- Produced 60+ content assets per month with just three people
- Recorded zero QA-related incidents
- Achieved 100% on-time delivery with zero escalations

Now, Stefan Milicevic, Strategy Director at Underground Ecom and Klaviyo Community Champion, is sharing the approach behind those results.

Stefan is a retention and CRM strategist with a decade of experience helping brands improve lifecycle marketing, customer experiences, and revenue. At Underground Ecom, he develops scalable retention strategies that combine data, automation, and AI with hands-on marketing expertise.

In this exclusive Customer Showcase, Stefan will take you behind the scenes of his team's AI-assisted quality assurance process and show how he uses Klaviyo Composer to investigate real marketing performance challenges.

What you'll learn

1. How to scale campaign production without sacrificing quality. See how Stefan's team built an AI-assisted QA process to review high-volume email campaigns and flows. Learn how they compare rendered test emails against approved content, flag personalization errors, missing disclaimers, incorrect placeholders, and other potential mistakes, and organize the findings for human review.

2. What to automate with AI, and what to keep human. Learn how Stefan separates repetitive QA checks from decisions that require human judgment, including visual quality, live-link checks, targeting, and final approvals. Discover practical ways to apply a similar approach to your own marketing workflow.

3. How to use Composer to investigate a drop in flow performance. See how Stefan uses Composer to analyze an abandoned-cart flow that experienced a 9% decline in deliveries and an 8% drop in conversion value, even as click and conversion rates remained steady. He'll walk through how he examines email-level performance, identifies where the decline is concentrated, and determines what to investigate next, from traffic sources to send delays and promotional offers.

4. How to get more useful, actionable results from AI. Learn how Stefan provides AI with reliable inputs, asks for structured, evidence-backed answers, and uses the findings to inform marketing decisions rather than accepting recommendations at face value.

Already using Composer? Pick up ideas for troubleshooting performance and improving how you prompt.

Haven't tried Composer yet? See practical examples of what it can do and where to start.

We'll leave time for live Q&A, so bring your questions about AI, marketing operations, or getting more value from Composer.