SaaS OmniSupport agent

Write the prompt

Start by writing a clear initial prompt that defines the purpose of the bot.

You are OmniSupport, an inbound customer support voice assistant for
OmniDimension. Speak in a friendly, polite, and professional tone.
Your main goals are to:
1. Greet the caller and introduce yourself as OmniSupport.
2. Understand the caller's intent. Whether they want to learn about the
   product, explore features, or book a demo.
3. Provide brief, clear responses about OmniDimension's offerings:
   AI-powered chat and voice agents, workflow automation, integrations.
4. If they request a demo, collect their name, email, and preferred time,
   and confirm the booking.
5. Handle unclear queries with a helpful fallback, asking the user to
   rephrase.
Always end the call with a thank you and a warm closing message.
Website URL: https://www.omnidim.io/

Web scraping

Automatically scrape the website and create a knowledge base from the URL provided in the prompt.

• The bot will automatically extract information from the website
• Create a structured knowledge base for the agent
• Attach the knowledge base to the bot for real-time information retrieval

Cal.com integration

Integrate Cal.com for sales or demo call booking.

• Navigate to the Integrations tab
• Click Connect next to Cal.com
• Complete the integration steps
• Once connected, the calendar will be automatically attached to the agent

See the Cal.com integration guide for the full walkthrough.

Slack integration for post-call

Integrate Slack for post-call delivery.

• Navigate to the Integrations tab
• Click Connect next to Slack
• Complete the integration steps
• Navigate to the Post-Call tab
• Choose Slack as the delivery method
• Choose the channel where you want to send the message
• Configure the details to send (full conversation, summary, extracted variables, etc.)

See Slack integration and post-call actions for full details.

Add a Make.com webhook

Add a Make.com webhook link and configure the scenario to update an Airtable database.

• Create a webhook in Make.com
• Set the webhook URL in your post-call agent configuration
• Test the webhook by sending a sample payload

See Make, Zapier, n8n, GHL and post-call actions for the full configuration reference.

Final testing and monitoring

• Re-test the agent via chat and web call to confirm all functionality
• Validate the booking flow, fallback responses, and knowledge base retrieval
• Navigate to the Call Logs section to monitor call interactions with evaluation details