Every new patient gets the same clear welcome pack: where to park, what to bring, how long the first visit runs, what happens afterwards, and who to call. It goes out automatically when a first appointment is booked, and it answers the ten questions your front desk currently answers by phone, one caller at a time.
Start from the Team Wiki Portal. It is a live Taskade Genesis app that answers questions from the content you give it, which is exactly what a welcome packet has to do, though its sample content is not yours. Ask it something, then clone it into your own workspace in about ten seconds and load your own practice details in place of the examples.
The build gives you the packet itself in short readable sections, a searchable version people can ask questions of, per-appointment-type variations so a first consultation and a routine visit get the right details, a record of who received which version, and an AI agent that drafts each section from your existing website copy, signage, and phone notes rather than from a blank page. It also flags gaps, such as a parking section you have never actually written down.
The packet is usable in several shapes:
- List view as an ordered pack people read top to bottom
- Table view to track who received which version and when
- Board view to move draft sections through review and publish
- Plus the rest of the 7 project views
Automations send it at the right moment. On paid plans, Taskade automations fire the packet when a first appointment is created, resend on request, and update anyone booked in future when a detail changes, such as new parking arrangements. Across 100+ bidirectional integrations, your calendar pulls the booking in while Gmail pushes the packet out.
The AI agents carry 34 built-in tools including web search and persistent memory. The agent keeps the packet consistent with what your site says, so a phone number changed in one place does not survive in three others.
Workspace DNA keeps the packet true. Memory holds your practice details, your existing wording, and every question people have asked that the packet did not answer. Intelligence, drawing on 15+ frontier models from OpenAI, Anthropic, Google, and open-weight providers, drafts each section from those sources and flags the gaps. Execution sends the right version on booking and notifies everyone affected when a detail changes.
Running it in your practice
- Collect what you already say. Website copy, printed sheets, and the answers your front desk repeats on the phone. That is the raw material.
- Clone the app and let the agent draft. The first output is a complete packet plus a list of things you have never actually written down.
- Fill the gaps, then cut. Practices almost always find the parking and arrival section is the one that was missing and the one people most needed.
- Send it on first booking. On a paid plan this is automatic, so the packet arrives while the appointment is still fresh in mind.
- Read the unanswered questions monthly. Whatever the searchable version could not answer is next month's new section.
The phone gets quieter within a fortnight, and the calls that remain are the ones that genuinely need a person.
Clone it, invite your team, and the words are yours. Most practices edit the first draft in an afternoon and then only touch it when something real changes.
Go deeper: build a customer portal if you want a signed-in home for the packet, the Community Gallery for published examples, or templates for a starting structure. This pairs with the pre visit prep checklist for the do-this-before part, and with the new patient journey map when you want to see where the packet sits in the whole experience.
Frequently Asked Questions
Where does the content come from?
From what you already have: your website, your printed sheets, and the answers your front desk gives on the phone. The agent drafts from those sources and flags anything it could not find, which is usually the most useful output of the first run.
Can different appointment types get different packets?
Yes. One packet with conditional sections covers most practices, and the appointment type on the record selects which sections appear. Nobody reads instructions that do not apply to them.
How do people ask a question the packet does not cover?
The searchable version answers from your content and says plainly when it does not know, rather than inventing something. Unanswered questions are logged, which tells you what to add next.
What if a detail changes after we have sent it?
Update the section once. On a paid plan an automation can notify everyone booked in the future that a detail changed, so nobody turns up at the old entrance.
Does the packet contain any clinical information?
Only what your practice writes and approves. The app organises and delivers your content. It does not generate guidance about care or add anything you did not supply.
Can we brand it with our own look and domain?
Styling is available on any plan. Publishing on your own domain with sign-in is a Business plan feature and above, and App Users is in beta. See custom domains.
How do we know people actually read it?
The record shows who received it and when, and the searchable version logs what people asked. Together those tell you which sections earn their place and which get skipped.
