Introducing ElevenAgents Architect. It's an expert built into ElevenAgents that helps your team launch and improve AI agents, just by talking or typing.

Ask ElevenAgents Architect: - Why customers asked for a human agent on refund calls - How to improve resolution rates on account queries - To build a test set that keeps your agent on brand It knows best practices, analyzes your transcripts, then suggests and implements improvements.

ElevenAgents Architect can also reason across thousands of your conversations to spot opportunities you’d otherwise miss. Each one comes back as a proposal, with the fix already built and validated in simulation.

Need a new agent? Just describe the agent you need and attach files for context. ElevenAgents Architect builds it the way an expert would, from the prompt to the guardrails to the knowledge base.

You can control how it builds so it’s aligned with your team’s way of working. And every change lands as a versioned draft you can review and revert, so nothing goes live without your approval and you can ensure changes are tested in simulation and piloted before a full rollout.

ElevenAgents Architect lets any member of your team manage agents wherever they work. Talk to it in ElevenAgents, or reach it from Claude, Claude Code, ChatGPT, Cursor and Grok Bot.

Learn more:

the useful part is having AI help build the agent instead of just answering questions about it

like that Architect doesn’t just suggest improvements , it builds and validates the fix in simulation.

an agent that builds agents, love it real test for me: can i just describe my voice agent in Hinglish and it gets it? that's how most Indian teams will actually talk to it

Talk to build agents? That’s huge.

Which model powers ElevenAgents Architect in the dashboard? Is it an ElevenLabs model or a third-party LLM, and can users choose which model it uses?

how does the conversation feature handle complex requests from users?

what kinds of improvements can Architect suggest - does it analyze conversation logs to identify where agents fail, or does it tune prompts and tool routing based on your team's feedback?

Talking the agent into a better prompt is fine until the latency and turn-taking still fall apart on a real phone call.

Launching and improving AI agents just by talking or typing ElevenAgents Architect makes agent building actually conversational!

Wah, keren banget nih ElevenAgents Architect! Bisa bikin tim makin jago ngembangin AI agent cuma modal ngobrol doang. Semoga fitur ini bantu banyak developer ya.

The agent gent that builds agents. Middle mgmt but make it voice.

an architect agent built into the platform will make voice agents far easier to ship

the second comment is the whole point honestly. everyone obsesses over agent speed but the real money is knowing why someone gave up on your bot and called a human instead. thats the feedback loop that actually fixes things.

Building AI agents just got much easier.

AI agent building just got a lot easier.

Love this direction making building assistants truly effortless

That sounds pretty cool.

"The age of Agentic AI is officially here. Lowering the barrier to building AI agents is exactly what the industry needs to accelerate business automation. Incredible work, ElevenLabs!"

Building agents by talking to an expert is a smart way to lower the setup barrier. The hard part after launch is testing them against real customer conversations.

Impressive!

This makes building AI agents feel much more accessible. Being able to launch and improve an agent just by explaining what you want could remove a lot of the technical friction for teams.

That sounds like a game changer for teams working with AI. Really impressive stuff.

Smart addition making agent building more accessible

This is a big step toward making AI agent development more accessible. 🚀 Having an expert that can help design, launch, and improve agents through simple conversations removes a lot of the complexity for teams. The future of building agents feels more collaborative. 🤖

Turning agent building into a conversational workflow is a big step toward making AI development more accessible.

The sounds its so mmm 🤌

That sounds like a game changer. Making AI agents easier to us

An agent that reads your own transcripts and tells you why callers asked for a human is the right tool. That's where the real fixes are. My take on why voice beats chat for support:

Dm for paid colloboration

Interested your post