The Pre-Session Questionnaire That Used to Travel as a Word File

The first version of this feature started with a Word document.

A family counsellor we work with sent every new client a pre-session questionnaire before the first appointment. It was a good questionnaire, written over years of practice. It asked about the client's health, their relationship, their family history, and in some cases about abuse and pregnancy losses. It went out as a .docx over WhatsApp, the client filled it in on their phone as best they could, and sent it back the same way.

Think about where that file ends up. In two WhatsApp histories, in the phone's downloads folder, in a cloud backup, possibly forwarded to a partner who "just wants to help fill it in". Those are some of the most sensitive answers a person ever writes down, and they were travelling as an ordinary attachment.

What we built instead

A survey in TalkOmni is a form whose answers go somewhere else. A normal form becomes a lead that your team can see and export. A survey's answers are encrypted with their own key the moment they are saved. Only the company owner and company admins can open them. The assistant cannot read them at all: no tool, no search, no summary mail reaches a survey answer, and there is a test in our code that fails if one ever does. They never go into an Excel export.

The person accepts a consent text before they see a single question, and the version of the text they accepted is stored with their answers. The answers are deleted a set number of days after the appointment, 90 by default. And when someone asks to be forgotten, one erasure request removes their responses along with everything else we hold about them.

A few practical things turned out to matter more than we expected:

• Saving page by page. A long questionnaire is not filled in at one sitting, especially by a parent at 11 at night. Every page is saved, and the same link opens where they stopped.
• A table field. The counsellor's form had a "life timeline": age, date, what happened, how it felt. As many rows as the person needs. Before, that meant forty separate boxes. Now it is one field where the client adds rows.
• Rules. A couples questionnaire and an individual one share most questions. A rule opens the "your relationship" section only when the session type is couple, so nobody scrolls past questions that are not for them.

Sending it without anyone sending it

The survey can be ticked to go out after every booking. When the assistant books a session in a chat, or the client books through the form on the website, they receive a personal link on the channel they used, and by e-mail if we have their address. The link works until the day after the appointment. The counsellor sees on each appointment whether the survey was sent, started or completed, and opens the answers right beside it.

The mistake that cost us

For the first week the project had one survey, and it was the one for children: 112 questions, written for parents booking a session about their child.

Every booking got it. Adults booking a session for themselves opened a link and found 112 questions about "your child's development". On 22 September we looked at the numbers and the completion rate was 16.7 percent. Most people who opened it simply closed it again, which is exactly what we would have done.

The fix was a question on the booking form, "Who is the session for?", and a small setting that maps each answer to a survey. Adults get the individual questionnaire, parents get the children's one, couples get the couples one. By 28 September, 17 of the 20 survey responses tied to an appointment had come through a routed survey. It is too early for a new completion rate we would trust, but people are at least being sent questions that are about them.

We should have caught this before the first booking. A default that is right for one kind of client is wrong for all the others, and the only person who notices is the client, who says nothing and closes the tab.

One honest limit

The questionnaire answers are encrypted field by field. The chat conversations are not. They sit in a database that is encrypted at rest, which is the normal standard for ordinary personal data, but it is not the same protection. So if a client starts writing clinical detail into the chat, the right move is to steer them to the questionnaire. We say this to every practitioner before they start, and your own lawyer should look at how you collect explicit consent for health data. That part stays with you as the data controller.

If your pre-session form is still a Word file today, the change is small: rebuild it once in the designer, tick "send after booking", and the file stops travelling.