1
Build an agent
Create an agent and give it a persona and prompt, a voice, and a
language. This is where you decide how it sounds and how it behaves.Lily, your Catalyst, Vocily’s agent assistant, can do this with you: she creates the agent, interviews
you about what it is for, and writes the prompt for you to review.Building agents → · Lily →
2
Add knowledge and abilities
Attach knowledge bases so the agent answers from your own content, and add tools or your own APIs so it can take real actions - before and during a conversation.Tools & knowledge → · Integrations →
3
Deploy to a channel
Put the agent live: make phone calls, connect it to WhatsApp, or embed a
Web Talk or Web Chat widget on your site.Phone Calls → · WhatsApp → · Web Talk → · Web Chat →
4
Have conversations
Your agent handles conversations on its own - following its instructions, drawing on its
knowledge, using its tools, and handing off to a human when it should.
5
Reach out at scale
Beyond inbound, run outbound call campaigns over the phone and WhatsApp campaigns
to reach many people at once - paced and tracked automatically.Outbound phone setup → · WhatsApp Campaigns →
6
Review and improve
Every conversation is captured. Read transcripts, listen to recordings, and see
summaries, sentiment, and custom analysis - then refine the agent and repeat.Conversations & analytics →
A few terms you’ll see everywhere
These are the building blocks the rest of the docs assume you know. The
The Account and workspace section covers workspace ownership,
access, and the organization model in depth.
- Agent - the AI that talks to your customers. One agent can be configured once and deployed across phone calls, WhatsApp, Web Talk, and Web Chat.
- Lily, your Catalyst - Vocily’s agent assistant, usually just Lily. She can build an agent, rewrite its prompt, and change its settings - always as a change you review before it is written.
- Channel - where an agent is deployed: phone calls, WhatsApp, Web Talk, or Web Chat.
- Conversation - a single call or chat session, with its transcript, recording, and analysis.
- STT - speech-to-text. It converts what the customer says into text the agent can understand.
- TTS - text-to-speech. It turns the agent’s response into spoken audio.
- Voice - the selected TTS voice and voice settings, such as speaking speed, that determine how the agent sounds.
- Knowledge base - a collection of your content the agent can draw answers from.
- Workspace - where your agents, numbers, contacts, and conversations live. Your account can have one or several.
Ready to build?
Follow the quickstart to create your first working agent.