Klaviyo has embedded two AI agents—Composer and Customer Agent—into a unified CRM platform designed to eliminate the fragmentation that typically separates marketing and service operations. Composer, now in public beta, identifies revenue opportunities by auditing live campaigns and segments, then builds targeted email and SMS campaigns with human approval gates. Customer Agent, already available as a full platform, handles service interactions whilst maintaining visibility into purchase history, marketing engagement, and customer intent signals. The critical distinction here is architectural: both agents operate from the same real-time customer profile rather than pulling data from siloed systems. This shared context means that when Customer Agent resolves a support ticket, it writes preference and intent signals back to the customer record that Composer can immediately activate in future campaigns. Conversely, campaign engagement data informs how Customer Agent personalizes subsequent service interactions, creating a feedback loop between revenue and support activity.
The implications for CX teams are substantial but require honest assessment. For organisations already running fragmented stacks—Zendesk for support, separate marketing automation, disconnected analytics—Klaviyo's approach exposes a genuine operational liability: your agents (human and AI) are working from incomplete customer pictures, and your teams are duplicating context-gathering work. The question becomes whether your current vendor ecosystem can deliver this kind of bidirectional data flow, or whether you're locked into unidirectional integrations that treat service as a cost centre rather than a revenue signal. For teams considering platform consolidation, Klaviyo's model demonstrates that the next competitive advantage lies not in having more AI tools, but in having AI tools that share memory and can act on it. However, this also raises a harder question: if shared customer context becomes table stakes, what happens to best-of-breed vendors that excel at service or marketing but lack the CRM infrastructure to anchor multiple agents in a single customer record?
The broader shift here reflects a maturation in how CX leaders should evaluate AI investments. Standalone chatbots and marketing automation agents have proven insufficient because they recreate the departmental silos that frustrate customers in the first place. Klaviyo's framing—that customer context is now a revenue issue, not just an operational one—signals that support interactions should be treated as commercial intelligence opportunities. This means your support team's work directly influences campaign performance, and your marketing team's actions shape support complexity. For administrators and team leads, this demands a fundamental rethink of how you measure success: support resolution rates alone no longer capture value if service interactions aren't feeding intent signals back into the revenue engine. The risk is that organisations without this integrated approach will find themselves increasingly disadvantaged as competitors use service data to drive more precise, timely marketing interventions.
Klaviyo Embeds AI Agents in its CRM to Deliver Shared Customer Context CX Today