How do I set up custom actions in Chatref for CRM support?

Real estate teams handle a flood of repeat inquiries: property inquiries, showing requests, new buyer registrations. Manually copying chat details into your CRM wastes time and introduces errors. With CRM custom actions for real estate, your Chatref agent captures the details in-conversation and pushes them to your CRM instantly. That means lead records are accurate, follow-ups trigger on time, and your agents spend time closing, not typing. It’s how you automate CRM tasks without sacrificing the personal touch of a human conversation.

Open the agent builder for the bot you want to extend. Navigate to Custom Actions. Click Add action, then:

1. Name the action clearly (e.g., “Create lead in Follow Up Boss”).
2. Define a trigger phrase - the natural language your visitors will say, like “I want to see this house” or “register me.”
3. Add fields the agent should collect - buyer name, email, phone, property MLS ID, desired showing time.
4. Set your webhook URL - the endpoint of your CRM (e.g., a Zapier webhook, built-in CRM API, or custom middleware). Chatref sends a JSON payload with the collected fields when the action fires.
5. Test the action right in the agent preview sidebar. Chatref confirms the payload structure before you publish.

That’s it. No code, no YAML, no manual mapping you have to babysit. Chatref automatically handles the handshake between the chat and your CRM.

Once the custom action is live, CRM action triggers work in two directions. The user says the trigger phrase, the agent collects required data, and Chatref fires the webhook. Your CRM receives structured data you can use to create or update records, assign a drip campaign, or log an activity.

You control what the agent says in response - a simple confirmation, a booking link, or a handoff to a human. The website-widget displays this entire flow inside your site, so the visitor never leaves the page. Customize Chatref for CRM by adding multiple actions: one for registrations, another for showing requests, a third for offer inquiries. Each runs independently, all from the same agent.

• Keep the trigger phrasing natural. Use the exact language buyers and sellers actually type, not internal CRM jargon.
• Collect only what you need. Ask for name, phone, and property. Fewer fields means higher completion rates.
• Map optional fields intelligently. If you have an MLS ID field, prepend the listing’s URL in the confirmation so the lead can verify it.
• Combine with human handoff. For complex negotiations, let the custom action collect info, create the CRM record, then route to a live agent.
• Review the insights summary. Chatref tags conversations that fired each action, so you can see which triggers convert and refine them.

How to create custom actions in Chatref for CRM support?

Open your agent in the Chatref app, go to Custom Actions, click Add action, define a trigger phrase, add fields for the data you need (name, email, property info), and set the webhook URL where the payload should be sent. Save and test from the agent preview. No code required.

What are the common use cases for CRM custom actions?

For real estate, the most common are: registering new buyer leads, booking showing appointments, requesting property valuations, logging offer details, and triggering follow-up drip sequences. Essentially, any repetitive data-entry task you currently do manually after a chat.

Can I automate CRM tasks with Chatref?

Yes. Custom actions let your AI agent push structured data directly to your CRM via webhooks. This automates CRM tasks like creating contacts, updating deal stages, scheduling activities, and logging notes - all triggered by the visitor’s chat input, no manual handoffs.

How to integrate custom actions with my CRM system?

Chatref sends a JSON payload to any public webhook URL. Use your CRM’s native API, middleware like Zapier or Make, or a custom endpoint. Provide the webhook URL in the custom action configuration. Test the integration in the agent builder to confirm your CRM receives the data correctly before deploying the widget to your site.