How to connect participant onboarding what to expect help…

Your participant-facing content is the source. That includes every document, page, or note that explains the trial journey: screening steps, what to bring, consent details, visit schedules, compensation rules, and contact information.

Chatref’s knowledge base ingests that content – PDFs, website URLs, plain text – and turns it into a grounded, searchable set of facts. The AI agent answers participant questions solely from that material.

The website widget is how those answers reach participants. A single JavaScript snippet on your trial site lets anyone ask questions and get answers in a chat interface, drawn directly from the knowledge base you built. So the connection is: your content → Chatref knowledge base → widget on your site.

1. Gather all participant onboarding materials. Pull together every PDF, FAQ page, and information sheet that explains what participants can expect – screening logistics, consent explanation, first-visit procedures, preparation instructions, and who to contact with problems. The more complete the set, the better the answers.

2. Add the content to Chatref. Log into app.chatref.ai, create an agent, and upload those files or point it at the relevant pages on your existing website. Chatref reads each source and makes it available for answering questions.

3. Review the agent’s behaviour. Optionally set a name and a friendly brand colour so the widget feels like your site, and tweak the default greeting. Test the agent in the built-in playground with questions you expect – “How long does the first appointment take?”, “Can I eat beforehand?” – and verify the answers are accurate and clear.

4. Embed the widget. Go to the widget settings, copy the provided <script> snippet, and paste it just before the closing </body> tag on every page where you want the chat to appear. Make sure the domain of your trial site is added to the allowed list so the widget loads properly.

5. Test on the live site. Visit a page, click the chat bubble, and ask a few representative questions. Confirm the widget pulls the right information and that nothing in your site’s design breaks it.

A potential participant on your site notices a small chat bubble in the bottom corner (often branded with your chosen colour). Clicking it opens a conversation window.

The widget greets them and invites a question. When they type something like “What do I need to bring to the screening visit?”, the AI agent searches the knowledge base and replies with an answer that quotes or summarises the specific detail from your uploaded materials. The reply arrives in seconds, at any hour.

Participants can ask follow-ups, clarify, or change topics. The widget maintains context within the conversation, so they can go from “What happens during screening?” to “And what if my blood pressure is high?” without starting over. They never receive a generic internet search result; every answer is pulled from your own wording and clinical trial requirements.

If a question falls outside the uploaded content, the agent politely says it cannot answer and – depending on your setup – might suggest phoning the site’s main number. (Lead capture or human handoff features exist in Chatref but are separate from the basic knowledge base and widget; the core setup already handles the majority of routine onboarding queries.)

Widget does not appear. The most common cause is a domain mismatch. Check that the exact domain where you embedded the snippet (including subdomain) is listed in Chatref’s allowed origins. Also confirm the script tag is unmodified and placed immediately before the closing </body> tag, not inside a restricted container.

Answers are incomplete or off-target. This means the knowledge base is missing the relevant information or the source document is unclear. Reopen the agent in Chatref, add the missing content (maybe a PDF that explains the exact step), and re-test. If the source wording is ambiguous – e.g., “fasting required” without specifying for how long – the agent will struggle. Rewrite the source with precise details.

Widget loads but answers feel generic or overly formal. Review the uploaded content. If your “what to expect” documents use dense medical language, the agent will mirror that. Consider providing a short, plain-language summary alongside the original materials, or edit the documents to use the tone you want.

Participants ask questions that lead to dead ends. Use the conversation history (in Chatref’s inbox) to spot new topics that should be covered. Then add those topics to the knowledge base so the widget can answer them the next time.

What causes participant onboarding what to expect problems for Clinical Trial Sites & CROs?

Fragmented information is the root cause – PDFs stored on separate drives, staff giving slightly different answers by phone, and no central source that participants can check at any hour. When potential participants have to call during business hours, many drop out before they ever speak to someone. Sites that rely solely on a static FAQ page miss the chance to answer specific, worry-driven questions that change with each protocol.

How do I improve participant onboarding what to expect for Clinical Trial Sites & CROs?

Centralise all participant-facing information in a single knowledge source that can answer questions instantly. Use a tool that lets you update the content as protocols change, and measure the questions people actually ask so you can fill gaps. Embedding that knowledge base in a chat widget on your site gives participants a self-serve path – they get consistent, immediate answers and arrive at their first appointment better prepared and less anxious.