Signing up is only the first half of getting started. The screens that follow collect just enough to make the product feel like the user's own: a name, a photo, and the areas of wellbeing they want to focus on, then close on a confirmation rather than another form.

This is where deferring personalisation shows up most clearly. Each step asks for one thing at a time, so setup reads as a short sequence rather than a long registration, and the deeper preferences wait until someone has actually seen the product.

The wireframe treated this as a directory. A search field, then a flat list of therapists, with the assumption that everyone arriving here was starting from scratch.

The final design leads with continuity instead. If someone already has a therapist, the top of the screen is that existing relationship, showing the last session, the next one and the agenda, and browsing moves below it. Most people opening this screen are returning, not choosing for the first time.

The list itself gained what people actually decide on: specialism, rating, and how many reviews it rests on. Search and filters stayed, but as tools for someone who knows what they want rather than the default state.

Search assumes someone already knows what they are looking for. Filters are for everyone else, and they carry the criteria people actually choose a therapist on rather than a generic keyword box.

Keeping them behind a control rather than on the page meant the therapy screen could stay calm by default, and still get out of the way quickly for someone with a specific need.

Sessions answers three questions in one place: what happened last time, what is coming, and how to book more.

The last session leads with a written summary rather than a recording, and keeps the recording one tap away. An upcoming session is a single dark card that expands into its detail, showing host, guests, reminder and agenda, so the screen stays quiet until someone needs it. Previous sessions run underneath, grouped by date.

Choosing a therapist is a high-stakes decision made by a stranger reading a page. The question this screen has to answer is not what someone offers, but whether they can be trusted with something difficult, and that changed the order of everything on it.

The wireframe opened with the booking action and put the reasons to book underneath. The final design inverts that: rate and rating together at the top, then the therapist's own words, then the things that can be checked independently, specialisms, certifications, years of experience, and reviews from people who have had sessions. Booking comes last, once the case has been made rather than before it.

Structuring it that way also kept the page factual. Nothing here argues that a therapist is good. It sets out what they are qualified in and what other people said, and leaves the reader to draw the conclusion.

The wireframe stacked every booking decision into one dense panel of dropdowns. It worked, but it read as paperwork.

The final design breaks the same decisions into four separate single-purpose rows, timezone, date, time and length, each already filled in with a sensible default so the common case is confirmation rather than entry.

The meeting agenda is the only free-text field and it moved to the end, after the mechanical decisions are done, when someone has the attention to write something considered.

The wireframe put messaging inside the therapist profile, reachable only once you had opened someone. The final design promotes it to the bottom navigation alongside Home, Therapy and Sessions, because a conversation already underway is something people return to directly rather than navigate back into.

That is also what keeps the rest of the product from growing sideways. Four destinations at the top level meant everything else could nest underneath without the navigation competing with the content.

The list itself is deliberately plain, just avatar, name, preview and unread count, because it is somewhere people pass through on the way to a conversation, not somewhere to spend time.