Klaviyo moved two artificial intelligence agents to public beta, embedding them directly into its business-to-consumer CRM platform so they can draw from the same real-time customer data. Brands including AS Beauty, Spanx and Dermalogica tested the agents - Composer, for marketing, and the Customer Agent - during a private beta that ended this week.
The move gives marketing teams on Klaviyo's platform automated tools that don't just execute tasks but learn from each other's work. Dozens of retailers in the Top 2000 Database already use Klaviyo as their marketing automation platform. Those retailers made more than $9 billion in 2025 ecommerce sales, according to the market research tool that tracks North America's largest online retailers by annual ecommerce sales.
How the two agents feed each other
When the Customer Agent resolves a conversation, it records preferences, product interest and intent signals. Composer then pulls that record to build smarter campaigns. When Composer launches a campaign, the resulting engagement data informs how Customer Agent personalizes its next interaction. The more each agent works, the smarter both become.
"Businesses aren't struggling because they lack AI tools - they're struggling because most AI can't act on the context that matters," said Jamie Domenici, chief marketing officer at Klaviyo. "The next era of consumer experiences will be defined by companies that can combine AI with deep customer understanding to move faster, personalize better, and turn customer insights into growth."
This feedback loop between service and marketing is central to Klaviyo's approach. Rather than operating in separate silos, both AI Agents & Automation tools share a unified view of each customer, something Klaviyo argues most bolt-on AI tools cannot do.
What Composer does for marketing teams
Composer runs full audits of every live campaign, flow and segment, then surfaces a ranked list of revenue opportunities. It might flag an abandoned cart flow that hasn't been updated in months, a welcome journey where new customers drop off before purchasing, or a lapsed segment of high-value buyers a brand hasn't contacted.
Once a marketer selects an opportunity, Composer recommends specific improvements and builds a cross-channel campaign to capture it. The agent draws on 14 years of customer context, best practices and patterns from nearly 200,000 brands, plus the brand's own campaign performance and signals. Klaviyo said it built the tool to solve two problems: knowing where to focus and not having enough time.
Katherine Cabe, senior director of retention marketing at AS Beauty, said Composer provided "unbelievable" visibility into their marketing operations. "It didn't just identify issues, it helped us understand the context behind them," Cabe said. "Composer's Flow Audit tool surfaced collision issues across our 113-flow library, showing us where automations were unintentionally competing for the same customers, creating unnecessary message volume, and potential revenue loss."
Customer Agent runs across channels from day one
Customer Agent comes pre-trained on a brand's data, voice and policies, which Klaviyo said eliminates the typical engineering setup time. It uses pre-built connectors and APIs to act - not just describe. Instead of explaining a return policy, it completes the return. Instead of describing loyalty points, it applies them.
Brands configure the agent once and it runs across web chat, email, SMS and WhatsApp. Klaviyo said brands can also embed it in a custom surface. A single prompt can create an entire service experience: Klaviyo gave the example of a running brand describing a "Find My Perfect Gear" quiz, with the AI building the questions, logic and product recommendations automatically. After going live, the agent checks its own performance daily and identifies what isn't resolving.
For AI for Marketing teams, the shift toward agents that share data across functions marks a practical step away from isolated automation tools. Both Composer and Customer Agent pull from the same customer records, so campaign performance directly shapes service interactions and vice versa.
Why this matters for marketing professionals
Marketers who use Klaviyo now have agents that can audit their entire campaign library, spot revenue leaks like flow collisions, and build cross-channel fixes without manual setup. The agents share a customer record, so service interactions feed campaign logic and campaign engagement data feeds service personalization. For teams managing dozens or hundreds of automated flows, the audit capability alone could surface problems that would take weeks to find manually. The public beta means these tools are available to any brand on the platform, not just early-access partners.
Your membership also unlocks:
