AI Agent Creation Guide

01 — Open AI Agents Manager

Log in to the Fonada Console and open AI Agents from the left menu. The AI Agents Manager lets you create, manage, and deploy agents across all channels.

1. Log in at fonada.ai.
2. Click AI Agents in the left-side menu.
3. The AI Agents Manager opens (fonada.ai/dashboard/agents).
4. Click Create Agent (top-right).

Click image to enlarge

Top navigation tabs

Page actions

• Create Agent — opens Agent Builder (top-right).
• Manage Tools — configure platform tools.
• Search agents by name — filter the agent list.
• Filter by Channel — All Channels dropdown.
• Showing X of X — total agent count.

Agent card information

• Status badge — Active (green) or Inactive.
• Channel badge — e.g. Chat Bot.
• Direction badge — INBOUND / OUTBOUND / FLOW.
• Agent name, description, created/updated dates, Agent ID.
• Quick actions: Edit, Analytics, Delete, Widget Code, Webhooks, Active toggle.

02 — Select deployment channel

After Create Agent, the Agent Builder shows Choose Your Channel — select where you want to deploy your AI agent.

How to select a channel

1. Click the desired channel card — a green border and checkmark (✓) appears.
2. Click Continue to Basic Info (blue button below the grid).

Click image to enlarge

03 — Configure basic information

After channel selection, the 4-step flow begins: Basic Info → Channel Setup → AI Config → Tools & Events.

Agent configuration progress bar

Description templates available

• Customer Support Bot
• Sales Assistant
• Appointment Scheduler
• FAQ Bot
• Onboarding Guide
• Custom Description (recommended for full control)

Agent conversation mode

LLM-based Agent (recommended) — AI-powered conversational model using Large Language Models.

Proceeding to next step

1. Fill all required fields (marked with *).
2. Click Next Step (bottom-right).
3. Use Previous (bottom-left) to go back.

Click image to enlarge

04 — Channel setup

The progress bar advances to Channel Setup (Step 2). For Chat Bot (Text Only), no additional configuration is required — the label shows No config needed.

Click image to enlarge

05 — AI configuration (prompts & LLM)

Configure the intelligence and behavior of your agent — the most important step in the wizard.

• LLM Provider * — e.g. Fonadalabs.
• LLM Model * — e.g. 3.5 / 35B.
• System Prompt * — behavior and personality (max 5000 words).
• LLM Temperature — randomness (lower = more focused).
• Token Limit — max tokens per response (e.g. 20000).

LLM provider options

LLM temperature guide

System prompt example

You are a professional customer support assistant.
Always answer politely and globally.
Provide accurate information.
If you do not know the answer, ask the user for more details.

Maximum supported length: 5,000 words

1. Fill LLM Provider, LLM Model, and System Prompt.
2. Set Temperature and Token Limit as needed.
3. Click Next Step to proceed to Tools & Events.

Click image to enlarge

06 — Tools & events (capabilities)

Final configuration step (Step 4 of 4). All prior steps show green checkmarks. Configure tool integrations and capabilities.

MCP platform tools

Use the search bar to find tools by name or click Refresh to reload the list.

How to enable a tool

1. Check the checkbox next to the desired tool.
2. The tool becomes active for your agent.
3. Use search to filter by name.
4. Click Refresh to reload from the platform.

Click image to enlarge

07 — Knowledge base configuration

After creation, open Dashboard → Agent-Builder. The active agent name (e.g. botname (chatbot)) appears in the top bar.

A green Connected indicator and Refresh Stats button appear on the panel.

Content input methods (tabs)

Adding text content

1. Click the Text tab.
2. Enter an optional title.
3. Paste or type content in the Content field.
4. Click Add to Knowledge Base (purple button).
5. Click Next Step to proceed to Deploy.

Click image to enlarge

08 — Deploy agent

Final stage when setup is complete. Channel badge (e.g. Chat Bot Text Only) and Active Agent: botname (chatbot) appear in the top bar.

Click image to enlarge

1. Review all configuration (use Previous if needed).
2. Click Deploy Agent (green, play icon).
3. Agent status changes to Active in AI Agents Manager.
4. Agent is live via Widget, Webhook, or API.

09 — Manage webhook API key

Open the Webhooks action on an agent card to integrate with external applications.

• Refresh Key — generates a new API key (invalidates the old one).
• Delete Key — permanently removes the current key.
• Manage API Keys — view all keys for the agent.

Sample cURL request

curl -X POST \
  https://${SUPABASE_PROJECT_REF}.functions.supabase.co/agent_chat \
  -H "x-api-key: agent_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "bot_id": "YOUR_BOT_ID",
    "message": "Hello",
    "session_id": "optional",
    "app_key": "optional"
  }'

Response body

{
  "reply": "Hello! How can I help you today?",
  "session_id": "generated-uuid-for-session",
  "llm_provider": "openai",
  "llm_model": "gpt-4o"
}

On the agent card, click Webhooks, then Create Key in the dialog to generate your API key.

Click image to enlarge

10 — Customize chat bot & embed widget

Use the customize icon on an agent card to open Customize Chat Bot — tabs Customize Bot and Quick Buttons, with a Live Preview panel on the right.

Bot branding

Widget style

Colors & font

Font Settings — configure font size and family for the chat widget.

Embed chat widget

From the agent card, click Get Widget Code to open Embed Chat Widget — copy the script, paste before </body>, download the script, or use Preview Widget.

Click image to enlarge

Customize appearance

Use the customize icon on the agent card to open branding, widget style, colors, and Live Preview.

Click image to enlarge

Saving customizations

1. Make changes in the Customize Bot tab.
2. Use Live Preview to see updates in real time.
3. Configure Quick Buttons for pre-set reply options.
4. Click Save Customization (green, full-width).
5. Optionally click Download Script for updated embed code.

Important notes & best practices

The Fonada AI Agent Builder provides a structured workflow for deploying AI-powered chatbots across multiple channels. By following the steps in this guide, you can configure intelligent agents for customer support, sales automation, and engagement.

For further assistance, click Docs in the Fonada Console header or visit fonada.ai for support resources.

— End of document — AI Agent Setup Documentation · Version 1.0