Klaviyo Composer: The AI Agent That Builds Full Campaigns From a Single Prompt

What Composer actually does

You type something like “Spring sale for lapsed customers who bought outerwear last year, 20% off, three touchpoints over two weeks” and Composer builds the whole thing. Audience segment, channel mix, send schedule, email copy, SMS copy, subject line variants for A/B testing. The whole campaign, ready to review.

That’s the pitch. And honestly, the first time I ran it, I sat there for a second just staring at what came out.

Composer went live on March 24, 2026. It’s trained on 14+ years of Klaviyo data across 193,000+ brands and billions of customer interactions. That’s not a marketing number — it’s what makes the timing and channel recommendations actually useful rather than generic. When it suggests sending the second touchpoint on Thursday morning, it’s drawing on patterns from brands in your vertical, not just a default schedule.

The numbers behind the AI features

Klaviyo has been publishing performance data on its AI features for a while now, and the numbers are worth knowing before you decide whether Composer is worth your time.

Automated flows built with AI generate 320% more revenue than batch-and-blast campaigns. AI-generated subject lines average +14% open rate improvement. AI product recommendations drive +35% more revenue per send. Personalized Send Time, which went GA on March 24 alongside Composer, shows +23% open rate improvement.

Those are aggregate numbers across a large customer base, so your mileage will vary. But they’re not cherry-picked case studies — they’re platform-wide averages.

Everything else that shipped March 24

Composer got the headline, but the March 24 release was bigger than one feature.

Customer Agent, Klaviyo’s AI support tool, expanded to handle returns and order editing, and added WhatsApp as a channel. If you’re running cross-border and your customers are on WhatsApp (which they are, in most of Europe and Southeast Asia), that’s a real addition.

Personalized Send Time moved from beta to general availability. It calculates the optimal send window per individual contact rather than per segment, which is a meaningful difference if your list spans multiple time zones.

RCS support also launched. RCS is the messaging standard that replaces SMS on Android with read receipts, typing indicators, and richer media. Adoption is still uneven, but it’s worth having in the channel mix now rather than scrambling to add it later.

The cross-border piece: Locale Aware Catalogs

This one shipped March 9, a couple weeks before Composer, but it’s directly relevant if you’re selling across markets.

Locale Aware Catalogs automatically syncs your Shopify Markets localized catalog into Klaviyo. That means product names, prices, and currencies in the right language for each market, without manual catalog management. Reebok Europe was one of the early users and cited it as the thing that finally made their multi-market email personalization manageable.

If you’re running separate Shopify stores per market, this doesn’t apply. But if you’re using Shopify Markets with a single store and multiple locales, it’s a genuine time-saver.

The price question

Klaviyo raised prices 25% in January 2026. That’s not a small increase, and it’s worth being direct about it.

For a store doing $2-5M in revenue with a 50,000-contact list, the monthly cost is now meaningfully higher than it was a year ago. The question is whether the new AI features justify the delta.

My honest read: if you’re using Klaviyo mostly for basic flows and occasional campaigns, the price increase is hard to justify. The AI features are genuinely useful, but they’re not magic — they still require someone who understands your brand and customers to review and adjust what comes out.

If you’re running complex multi-channel campaigns across multiple markets, the combination of Composer, Locale Aware Catalogs, and Personalized Send Time probably does save enough time to offset the cost. The math works differently depending on how much of your team’s time goes into campaign setup.

What Composer is good at

Composer is fast at the structural work: building the segment, picking channels, setting up the timing logic, generating first-draft copy. That’s the part that used to take a few hours of clicking around in the flow builder. Now it takes a few minutes.

The A/B test setup is genuinely useful. It’ll generate subject line variants automatically and set up the test structure, which is the kind of thing that often gets skipped because it’s tedious to configure manually.

For standard campaign types — win-back, post-purchase, seasonal sale — the output is solid. The copy is competent, the timing is reasonable, and the segment logic is usually right.

What still needs human judgment

Brand voice is the obvious one. Composer writes in a generic e-commerce register that’s fine but not distinctive. If your brand has a specific tone, you’ll be editing the copy.

The audience logic can be too broad. “Lapsed customers” is a reasonable starting point, but whether that means 60 days or 180 days, whether you exclude recent browsers, whether you want to split by purchase category — those decisions still require someone who knows your data.

Anything involving a new product launch, a brand story, or a campaign tied to a specific cultural moment needs human input. Composer doesn’t know what makes your product interesting. It knows what campaigns in your category tend to look like.

The output is a starting point, not a finished campaign. That’s not a criticism — it’s just accurate. The time savings are real, but they’re in the setup and structure, not in eliminating the need for review.

Worth it?

If you’re already on Klaviyo and running multi-channel campaigns, Composer is worth testing on your next campaign. The worst case is you spend 10 minutes reviewing and editing what it generates instead of building from scratch.

If you’re evaluating Klaviyo for the first time and the 25% price increase is a sticking point, the AI features are a real differentiator, but they’re not a reason to pay for a plan that’s otherwise too large for your current list size. Start with what you need and grow into the AI features.

Related Articles