How to set up knowledge base for after hours new client c…

You need a Chatref account –sign up here to get your free $50 credit, no card required. You’ll also need the practice material your new clients ask about most often after hours:

• A list of services and modalities (individual, group, telehealth, etc.)
• Insurance plans you accept and self-pay rates
• The intake process – what forms to complete, what to bring, how long the first session takes
• Common questions (confidentiality, cancellation policy, what to expect)

Everything you upload will power the AI’s answers, so include the same detail your front-desk staff would give over the phone. If you serve the Mental Health Services space, the kinds of files you already have (a patient FAQ sheet, a PDF of intake steps, your services page) are exactly what Chatref needs.

Make sure you know which details you want to capture from a new client at night – typically name, email, phone, and a short note about what they’re looking for. Lead capture collects those inside the conversation and stores them in Chatref for your team to follow up on in the morning.

1. Build your knowledge base

In your Chatref dashboard, create a new agent and upload your practice content. Use PDFs of your intake packet, copy-paste the services and insurance sections straight from your website, or point Chatref at your FAQ page. You can add multiple sources – the agent will read everything together.

For a mental health practice, make sure you cover:

• Services and approach – CBT, EMDR, medication management, etc.
• Eligibility and fees – insurance accepted vs. self-pay rates, sliding scale if offered
• New client pathway – exact steps from first inquiry to first appointment, including forms and any pre-visit tasks
• Logistics – hours, location, telehealth instructions, parking

The more nuanced your source material, the fewer incomplete answers you’ll see. Chatref will answer only with what you give it – no guessing.

2. Configure lead capture

Open the agent’s settings and turn on lead capture. Choose the fields you want:

• At minimum, name and email (to reach them next business day)
• Phone number if your practice calls back
• A free-text note field so visitors can describe why they’re reaching out

Set a trigger that feels natural – for example, after the visitor asks two questions or after the agent detects phrasing like “I’d like to book” or “How do I become a new patient?” Chatref will then gently ask for the details without interrupting the flow.

3. Customize and embed the widget

Pick a brand color, write a short welcome message like “Ask us anything about becoming a new client,” and grab the embed snippet. Place it on your website – typically on every page so the chat follows visitors regardless of where they land.

The widget is always active by default, so it naturally covers after-hours. You don’t need to set schedules; if no human is logged in, the AI answers on its own. When a lead is captured, the conversation sits in your Chatref inbox, complete with the captured fields, ready for your team to review.

Test the entire flow before you rely on it overnight.

1. Open your website from a fresh browser after hours (or use an incognito window).
2. Ask questions exactly how a prospective client would – “Do you take Aetna for therapy?” “What do I need before my first session?” “I’m looking for help with anxiety, how do I start?”
3. Pay attention to whether answers come back grounded in your own material. If you see an answer that feels generic, you’ll need to add more detail to your knowledge base.
4. When lead capture triggers, verify that the prompted fields appear naturally in the conversation and that the captured data shows up in your Chatref dashboard under the conversation.

Repeat with a few different question phrasings, including short, open-ended ones like “I want to see someone.” The agent should ask clarifying questions and then offer to capture contact info.

The agent didn’t capture the lead.
Check that lead capture is enabled on the correct agent and that your trigger condition isn’t too strict. A good rule of thumb: set it to fire after the second user message, not after a specific keyword.

Answers feel generic or wrong.
This almost always means your knowledge base is missing the detail the agent needed for that question. Add a short document – even a few paragraphs in plain text – that directly addresses the missing information. No need to retrain everything; the agent picks up new content automatically.

Lead capture fires too early, before the prospect gets their question answered.
Push the trigger further into the conversation, or set it to only appear after the agent has provided a certain number of answers. You want the visitor to feel helped before you ask for their details.

Visitors don’t start the chat.
Make sure your widget is visible and your welcome message is inviting. Something like “Ask us about starting therapy tonight – we’ll answer from your practice’s own details” can increase engagement. Also check that the widget loads correctly on mobile – most after-hours visitors are on their phone.

---

What causes after hours new client capture problems for Mental Health Services?

Three things most often: the chatbot doesn’t have enough practice-specific detail to answer intake-related questions, so visitors leave; the lead capture form pops up too early and interrupts the answer; or the practice’s website doesn’t make the chat widget obvious after hours. When any of these fail, you lose the inquiry before morning.

How do I improve after hours new client capture for Mental Health Services?

Start by uploading richer source material – add a plain-text document that explains the exact intake steps and what you accept. Set the lead capture trigger to fire only after the visitor has received at least one helpful answer. Then test the flow from a mobile device on your own site, outside business hours, and tweak the welcome message to orient visitors toward asking about therapy. Finally, review the conversation tags in Chatref to see what questions keep coming up and fill any remaining gaps.