From Static CRM to AI Marketing Agents With a Memory
Klaviyo’s Composer and Customer Agent are AI marketing agents embedded in a CRM platform that share a real-time customer profile, allowing service interactions to drive campaign automation and campaign engagement to inform future support, which shifts CRM from a passive record system into an active engine for autonomous campaign execution and customer context sharing.
This move to public beta is more than a feature drop; it is a line in the sand for CRM automation. Composer, Klaviyo’s AI marketing agent, now assembles cross-channel lifecycle campaigns from a plain-language brief, choosing audiences, drafting email and SMS copy, and suggesting send times. Customer Agent, expanded into an “agent platform,” runs on the same profile and event stream, resolving common service tickets and taking actions in downstream systems rather than only answering questions. In other words, Klaviyo is trying to turn what used to be separate tools and teams into a single, coordinated set of AI marketing agents that can see the same customer, in the same moment, and act without waiting for humans to wire everything together.
How Composer and Customer Agent Enable Autonomous Campaign Execution
The real shift is from AI that suggests to AI that does. Composer scans existing automations, spots underperforming flows, and surfaces “opportunities” to grow revenue. Once a marketer selects an opportunity, Composer recommends how to act and builds the full campaign across email and SMS—audience, content, and channel execution included. Composer is positioned as an AI marketing agent that can build a complete campaign from a plain-language brief, including audience selection, cross-channel copy drafts, and send timing.
On the service side, Customer Agent goes beyond scripted answers. Using pre-built connectors and APIs, it surfaces order data, policies, and loyalty status, then acts on them. Instead of explaining a return policy, the Customer Agent completes the return; instead of describing loyalty points, it applies them. It can do this across web chat, email, SMS, WhatsApp, and custom interfaces, backed by multilingual expansion to 100+ languages. This is autonomous campaign execution in practice: AI that not only writes copy or predicts churn but orchestrates lifecycle campaigns and closes service loops in near real time.
Customer Context Sharing: The New CRM Battleground
The most important design choice here is shared context. Both Composer and Customer Agent operate on the same real-time profile and event stream, so they can continuously update each other’s understanding of the customer. When Customer Agent resolves a conversation, it records preferences, product interests, and intent signals—size, color, categories—and writes them back to the profile. Composer then uses those signals to build smarter marketing campaigns and tighter lifecycle segments.
This two-way customer context sharing also works in reverse. When Composer launches a campaign, subsequent engagement—opens, clicks, replies—feeds how Customer Agent personalizes the next interaction. The value is not AI copywriting; it is an orchestration layer that audits what is running, identifies gaps, and builds the assets and steps needed to act, cutting down the operational burden on lean retention teams managing huge automation libraries. With retailers using Klaviyo driving more than $9 billion in ecommerce sales in 2025, even small efficiency gains in this orchestration layer can translate into meaningful revenue shifts.
Why This Signals a Broader Shift in CRM Automation
Klaviyo’s launch reflects a wider trend: CRM platforms are evolving into AI sales and marketing agents rather than staying as passive databases. AI agents are entering CRM, handoff, routing, and revenue workflows, and the pressure is now on vendors to deliver controlled automation—systems that not only generate content, but run processes end to end with human checkpoints, testing, and clear ROI reporting. As Klaviyo’s CMO put it, “Businesses aren’t struggling because they lack AI tools—they’re struggling because most AI can’t act on the context that matters.”
Composer and Customer Agent lean hard into that context, drawing on 14 years of customer patterns from nearly 200,000 brands to decide what to run and when. Composer public beta is available now, while Customer Agent is live as a full service platform with more features slated throughout 2026. This is not risk-free. Poor data rules can turn weak signals into growth risk, and teams still need QA, guardrails, and escalation paths. But the direction is clear: in the next wave of CRM automation, the winners will be those who treat AI as operational staff—autonomous marketing executors—rather than as a novelty widget strapped onto a contact database.
