How to set up custom actions for patient intake pre regis…

You need a Chatref account with an agent already trained on your practice information (hours, services, accepted plans, intake forms, pre-registration steps). Your website widget must be installed and live. If you are new to Chatref for healthcare, see the Hospitals & Medical Centers page for a full overview.

Also, map out the exact fields you must collect before a patient’s first visit. Typical ones: patient first and last name, date of birth, insurance carrier and member ID, chief complaint, preferred appointment time, and any pre-existing condition flags. Decide which are required and which are optional. Keep the list as short as possible - every extra field increases abandon rates.

Have your destination ready if you plan to forward the data. A Google Sheet, your scheduling system, or a simple webhook endpoint all work.

1. Go to Custom Actions
   In the Chatref dashboard, open the agent you use for patient intake. Click Custom Actions in the agent sidebar, then New Action.

2. Name the action
   Give it a clear label your team will recognise, like Pre-registration intake. The label appears in the chat widget when the action is offered.

3. Build the form fields
   Add fields one by one. For a standard pre-registration flow, include:

   • Patient First Name (short text, required)
   • Patient Last Name (short text, required)
   • Date of Birth (date picker, required)
   • Insurance Provider (dropdown or short text; populate it with the plans your practice accepts)
   • Member / Policy Number (short text, optional)
   • Chief Complaint / Reason for Visit (long text, required)
   • Preferred Appointment Window (short text, optional - e.g. “mornings”)
   • Phone Number (short text, required)

   Mark each required field so the form cannot be submitted with blanks. Use clear helper text under each field (e.g. “We will call you to confirm the time”).

4. Decide what happens after submission
   By default, collected data stays in the chat transcript for your team to review. If you want the data to flow into another system, configure a webhook. Paste a URL that can accept a POST request. Chatref sends a JSON payload with the field names and the patient’s answers. You can route it to your scheduling platform, a Zapier webhook, or a simple endpoint that writes to a spreadsheet. Test the webhook with a sample payload to confirm it works before activating the action.

5. Set when the action appears
   The agent needs to know when to offer the pre-registration action. The most reliable way is to include an instruction in the agent’s prompt or knowledge base, like:

   “When a new patient asks about scheduling or says they are a first-time patient, offer the Pre-registration intake custom action.”

   Alternatively, you can trigger it via a specific user phrase (e.g. “I’m a new patient”), but the prompt-based rule keeps it flexible. Save the action and the agent rules.

Open your website or the chat preview in the dashboard. As a new patient, type “I want to schedule as a new patient” or a similar phrase. The agent should greet you and then present the custom action form. Fill every field, submit, and verify:

• The filled fields appear in the chat transcript under the patient message.
• If you used a webhook, check your destination - the data should arrive within seconds, with field keys matching your form.
• Intentionally skip a required field and try to submit. The form should refuse to continue and point out the missing answer.

Test on a mobile device too. The widget should remain usable, and the date picker should open smoothly. Also test what happens when the patient starts a conversation but does not complete the action - the agent should keep the session alive and not lose any partial data (custom actions save only on submit).

The action never appears. The agent’s rules are not correctly instructing it. Add explicit instructions in the agent prompt and include example patient messages. Avoid relying only on phrase triggers unless the phrases are exact.

Webhook never fires, or returns errors. Verify the endpoint URL is correct and publicly reachable. Your server must accept the payload format Chatref sends (JSON, with action_id, field_values, and conversation_id). Check the Webhook Logs tab inside the agent settings to see the raw request and the server’s response.

Collected data is inconsistent or missing. Review the field configuration - required flags, drop-down options, and helper text. If you changed the fields after saving, the action may need to be re-published. Also, patients sometimes type free text that doesn’t match your system’s format; consider adding validation remarks in the helper text (e.g. “Use MM/DD/YYYY”).

Patients abandon the form. Too many required fields, unclear labels, or long text prompts cause drop-offs. Start with only the essentials (name, DOB, chief complaint) and add optional fields once the flow is proven.

What causes patient intake pre registration chat problems for Hospitals & Medical Centers?

The most common problems come from outdated practice information in the knowledge base - if your accepted insurance list or hours are wrong, the conversation breaks trust instantly. Poorly designed custom actions with too many required fields, unclear labels, or broken webhooks also fail. Another frequent issue is the agent not being prompted to offer the action at the right moment, leaving patients to type long free-text messages instead of filling a structured form.

How do I improve patient intake pre registration chat for Hospitals & Medical Centers?

Keep your knowledge base current with your latest forms, accepted plans, and any pre-visit instructions. Limit required fields to the bare minimum and test the form yourself before going live. Use simple, direct field labels (“Date of Birth” not “DOB”). If you forward data to another system, monitor the webhook logs frequently. Finally, review chat transcripts regularly to spot where patients get stuck, and tweak the agent’s prompt to guide them more smoothly into the action.