Behind the build
1 min

Most AI agents live in silos. Merlin builds across your business.

Tray.ai Co-founders Rich Waldron and Alistair Russell explain how teams use Merlin Agent Builder to create agents with built-in controls that act across systems.

Video thumbnail

What’s covered 

Most AI agents are stuck inside single apps—limited in scope, disconnected from context, and hard to govern. This short video breaks down how Merlin Agent Builder gives teams one platform to design agents that act across systems, use real-time context, and stay under control with built-in guardrails and audit trails.

Top takeaways   

  • Most AI agents are designed to chat—not to execute. Merlin powers agents that act

  • Design once, adapt fast—Merlin makes it easy to update agents as systems, workflows, and business needs evolve

  • Use context from past interactions and system data to guide smarter decisions

  • Governance isn’t an add-on: guardrails, audit logs, and human-in-the-loop are built in

Transcript

Most AI agents are built in individual SaaS apps. So they're isolated, they're siloed, they're hard to govern, and they're disconnected from the rest of your stack. This is a story we've seen many times over, fragmented apps that promise more than they can deliver. That's why we built the Merlin Agent Builder.

Merlin Agent Builder is an intuitive guided experience where you can build agents that work across your entire stack. It's not a one off. It's your central hub for building your agents.

You can start with one of our pre-built agents, or you can even build your own. You choose where your users interact with the agent, so through Slack or Teams, or even autonomously react to events across your system. And then the key part is you choose the systems where you want your agents to take action.

This is the only agent platform that combines real time action, governance, and integration in one place. So agents built with the Merlin Agent Builder can act with context from across your entire environment and act on it throughout your entire stack.

You get full audit trails. You can also define guardrails, which keep your agent on track, and you can ground the agent in your company's knowledge.

At any stage of the process, you can also introduce human-in-the-loop approvals.

The agents are going to learn from their context and interaction with users, and they're going to proactively solve what used to be very repetitive work.

With Tray, you're not locked into a single agent or a dead-end tool. You're building a foundation that actually grows with you. Finally, a single place to launch and scale your horizontal agents.

When AI agents can reason and act with context across your systems, they become truly useful. This is the power of Merlin Agent Builder.

Let's explore what's possible, together.

Contact us