Today, Klaviyo is launching Custom Skills for Customer Agent , which lets any brand build entirely new AI agent workflows for their specific use cases, in natural language and without engineers.
We know that the AI agent landscape has a control problem. VentureBeat research from Q1 2026 found that 72% of organizations claim two or more "primary" AI platforms, and most haven't built the governance, guardrails, or accountability structures to manage any of them well.
The gap between deploying AI and actually controlling it is widening. And in customer service, where the agent speaks as your brand, that gap becomes a liability.
Most AI service tools contribute to this problem. They give you a fixed set of capabilities behind a prompt editor. For the basics, like shipping questions, order status, or return policy lookups, that's fine.
But brands quickly hit the ceiling of what you can configure:
- Enforce a strict sequence of operations on a return.
- Verify identity before taking action.
- Route a gifting experience so the recipient's answers don't pollute the buyer's profile.
- Run a product-matching quiz, save the answers to the customer profile, and pick up the thread on the next visit.
You can't get to those outcomes with a prompt editor. And you can't get there by waiting for a vendor to build the capability for you. The more consequential the interaction, the more you need a system you can extend, not just one you can adjust.
That's the difference between a product and a platform. And it's why we built Customer Agent the way we did.
Start fast, with full flexibility
In March, we expanded Customer Agent with a full catalog of pre-built skills for retail: order tracking, returns, product recommendations, subscriptions, and loyalty. That catalog lets brands get live in days, not quarters.
But every brand has use cases the pre-built catalog doesn't cover.
- A fragrance brand where the highest-value question is "what should I buy?" — and the right answer depends on preferences the customer hasn't shared yet.
- A fashion brand whose gift-givers are some of their best customers, but whose answers should never touch their own customer profile.
- A restaurant group that needs a reservation assistant with location-level availability and a memory of your favorite table.
Custom Skills are how you handle those. You describe what you want the agent to do. You connect the systems it needs through the Tools. You set the guardrails including brand voice, escalation rules, and handoff behavior through Agent Guidance. And the agent goes to work autonomously, for every customer, every time.
The whole experience is designed for the person who knows the customer best—the marketer, the CX lead, the service director—not the person who writes the code.
Control your own destiny
Custom Skills, Custom Tools, and Agent Guidance aren't three features. They're three layers that, together, give brands the ability to control their own destiny on a platform designed for exactly that.
You don't have to wait for us to build the capability you need. You can build it yourself, with your policies, your catalog, your voice, and your way of doing things on infrastructure that makes this all possible.
In a landscape where most brands are assembling AI from multiple vendors with no clear governance model, that kind of control is rare. And it gets rarer as the stakes get higher. Every AI service interaction carries your brand's name. The ability to govern how an agent behaves—not just what it says, but what it does, what systems it touches, and what data it writes back—is the table stakes that most tools skip.
The data foundation that makes the platform work
Other AI service tools sit on top of helpdesks. Helpdesks manage tickets. They don't know the customer like Klaviyo does. They offer no details on that customer’s purchase history, loyalty tier, or browsing behavior, for instance.
When the conversation ends, nothing writes back to the customer profile. That means the next marketing campaign is no smarter than the last one.
Customer Agent is built directly on Klaviyo's data platform, where the customer's full history already lives. Before a customer sends a single message, the agent already knows who they are, what they've purchased, what they've browsed, and what they're worth to the business.
Every signal from every conversation writes back to the same profile that drives your marketing. Service makes marketing smarter. Marketing context makes service better. That feedback loop is the advantage that compounds.
And it's what makes the platform work. Every skill you build, every tool you connect, every interaction your agent handles, all of it reads from and writes back to the same customer profile. That shared context is what separates a platform you can build real experiences on from a product you can only configure.
What this means for your team
For marketers, Custom Skills open up a new kind of brand experience that only surfaces in conversation. The scent profile. The sizing frustration. The gift recipient. The fit preference. All of it becomes part of how your brand understands the individual across every future touchpoint.
For service and CX leaders, you can go beyond answering FAQs to offering personalized answers and guidance, as well actually automating resolutions end-to-end. And because every Custom Skill inherits the same Agent Guidance applied across every other workflow, everything you build automatically sounds like your brand.
Where this goes
For years, the aspiration in marketing has been personalization at scale. That meant the right message to the right person at the right time. Marketing technology has gotten very good at the message part with segmentation, A/B testing, and predictive send time.
But "the right person" has always been incomplete. Brands have only ever known half the customer—the marketing half, not the service half where customers tell you what didn't fit, what they love, what they're actually shopping for and why. That half lived in a helpdesk, disconnected from everything else the brand knew.
Customer Agent changes that. And Custom Skills is the piece that lets brands take real ownership of what their agent does, how it behaves, and what it learns because every interaction reads from and writes back to the same customer data that drives everything else.
Soon, you'll be able to simulate any skill against realistic customer scenarios before it goes live. The goal isn't just to launch an agent. It's to trust it enough to let it do more.
Custom Skills for Customer Agent are available now in managed beta. Join the waitlist.
Related content
