Marketing automation platform Klaviyo is “going headless”, opening its CRM platform to AI agents and allowing marketers to access and act on customer data from tools including ChatGPT and Claude without needing to open Klaviyo itself.
The move was announced at Klaviyo’s annual K customer conference in Boston today, alongside a raft of new AI, data and personalisation capabilities aimed at shifting the CRM from a platform marketers log into towards infrastructure that can work wherever they do.
Under the new headless architecture, more than 260 Klaviyo MCP tools and capabilities and more than 490 APIs can be accessed directly through Claude, ChatGPT or other AI systems.
That means an agent can read and write Klaviyo data and take action on a brand’s behalf without anyone opening the Klaviyo interface.
Klaviyo said brands are already using the architecture to build custom reporting dashboards with live Klaviyo data, automate weekly performance reviews and send recommendations directly to Slack.
Klaviyo co-founder and co-CEO Andrew Bialecki said AI agents need access to both customer data and the ability to act on it.
“An agent is only as useful as three things; whether it understands a brand’s data, whether it can reach that understanding from wherever it’s running, and whether it can act on it for every individual customer,” Bialecki said.
“Most platforms give a business one of those. With this launch, Klaviyo gives all three, built into one platform. That’s what separates real infrastructure from a walled garden.”
Klaviyo is also bringing SQL into its Klaviyo Data Platform, allowing marketers to ask questions about their customer data in plain English rather than relying on a data team to write technical queries.
Through Composer, MCP or another agent, a marketer can ask a question and Klaviyo will translate the request into SQL, run it against the brand’s data and return the answer alongside the query that produced it.
The feature is currently available as a preview through Composer and MCP, with queries able to access profiles, events, segments, lists and catalogues.
Klaviyo said its platform processes more than four billion signals a day, making the ability to interrogate that data without technical expertise an increasingly important part of its AI strategy.
The company is also replacing Marketing Analytics with a new Personalisation layer, designed to make individual customer decisions in real time.
Rather than using one large model to handle every personalisation decision, specialised models will determine who to reach, what to show them and when and where to reach them.
Klaviyo said the models will share customer context across channels while remaining configurable and transparent to marketers.
Meanwhile, Composer is gaining scheduled tasks, reusable “skills” and the ability to edit segments directly.
More than 138,000 Klaviyo users currently use Composer to analyse performance and create campaigns, according to the company.
The new functionality means marketers can schedule recurring jobs such as performance summaries or campaign checks, while teams can save their own workflows and brand guidelines so they do not have to repeat the same instructions every time.
