AI-powered notetaker for mental health professionals

An AI-assisted notetaking system accessible in both video and in-person sessions that records audio, generates transcripts, and converts them into structured clinical notes using industry-standard templates.

B2B2C

AI

Mental Health Tech

playspace.health

TEAM

Product Designer

Christine Chow

Product Managers

LA

MC

Developers & QA

NM

RM

Clinical Advisors

NS

SB

timeline

Sep – Oct 2025 (2 months)

Skills & tools

UX Research
UI/UX Design
Prototyping

Figma (Design / Make)
ChatGPT

Overview

At a Glance

Features

✂ - - - - - - - - - -

Research & Discovery

Key Insights

Design Iterations

Reflection

at a glance

Context

PlaySpace is an all-in-one platform for mental health professionals, combining practice management with interactive tools to deliver personalized and therapeutic activities for both in-person and virtual care.

Challenge

While preparing for broader market expansion, the team realized we needed more robust clinical documentation tools for practitioners who were drowning in administrative work in order to stay competitive.

According to a survey we sent out to 70 customers, 21.1% of respondents believed an AI-assisted notetaker would be helpful for their practice.

Impact

Created a custom design system

Using an open-source component library as a foundation, I delivered a usable lightweight design system in less than 2 months to ensure brand awareness and consistency across the platform.

View this project

Established Figma as our single source of truth

I onboarded the product and engineering teams to Figma; organized all design iterations. feedback, and comprehensive design documentation for a smooth handoff to developers; and recorded development inconsistencies to ensure design accuracy.

Lead regular interviews with clinical practitioners

As the sole product designer, I led weekly meetings with clinical advisors and regularly consulted the customer service team to validate designs with real workflows & pain points.

features

1 / 4

Multiple entry points for different workflows

We had identified two potential user types: practitioners using PlaySpace as their primary EHR, and practitioners using PlaySpace alongside their own primary EHR. Because of this, I considered multiple entry points so our note-taking tool could adapt to fit different practices.

2 / 4

Notetaking in virtual sessions

In a virtual session, practitioners can choose between taking notes without recording, or recording with quick notes — both options exist to give practitioners a sense of true ownership over their notes.

The quick notes interface is particularly designed to fit on the side, ensuring uninterrupted access to therapeutic tools and the client’s face.

During a recording, practitioners see a live indicator in the top bar while clients see a more subtle status.

3 / 4

From session to official documentation

After the session, default templates are automatically applied to help practitioners complete the official session note, with relevant supporting documentation available on the side for reference.

Industry standard templates (SOAP, DAP, BIRP) are offered out of the box, with the option to create custom templates tailored to individual preferences.

4 / 4

Signing & locking the official note

Incomplete session notes must be signed and locked with the practitioners’ signature, full name, credentials, and date and time.

Research & Discovery

While navigating an ambiguous opportunity space with emerging technologies, I led the discovery work to help define the right scope and priorities for the challenge. There were the basic objectives that were provided:

Support both in-person and virtual session recording without disrupting client care

Convert transcripts to clinical templates (SOAP, DAP, BIRP formats)

Ensure HIPAA compliance and legal documentation requirements

Over the course of several weeks, I conducted extensive research with support from the product team and facilitated regular user interviews with a practicing clinical advisor to understand:

  • Current documentation workflows and pain points

  • Legal and compliance requirements (HIPAA, consent, note retention)

  • Clinical template preferences and customization needs

I also took inspiration from competitors:

  • Klarify’s user flows solved key problems around AI-assisted clinical notes, so studying these patterns helped speed up our own exploration and dedicate more time to tailoring it to the PlaySpace experience.

  • Granola predominantly stood out from other AI notetakers for their user-first model, mixing users’ control over their own notes with AI as supplementary support rather than primary.

  • We also leaned on validated patterns from Owl and Jane — industry-standard electronic health record (EHRs) systems that therapists already trust for highly sensitive workflows like clinical documentation.

  • Klarify’s user flows solved key problems around AI-assisted clinical notes, so studying these patterns helped speed up our own exploration and dedicate more time to tailoring it to the PlaySpace experience.

  • Granola predominantly stood out from other AI notetakers for their user-first model, mixing users’ control over their own notes with AI as supplementary support rather than primary.

  • We also leaned on validated patterns from Owl and Jane — industry-standard electronic health record (EHRs) systems that therapists already trust for highly sensitive workflows like clinical documentation.

Key Insights

1

Two distinct user types emerged

Practitioners using PlaySpace as their primary EHR might have different workflows from those using PlaySpace alongside their own EHR. Thus, each group might require different entry points for documentation flexibility.

2

Practitioners value personal ownership over their notes

Every practitioner has their own method: sparse notes vs detailed client quotes vs industry-standard templates, session prep vs no prep at all. This variety revealed 3 main design needs:

  • We must provide industry-standard templates out of the box, as well as support customized templates to cover all bases.

  • We had to clearly distinguish between the types of notes (Session Prep, Quick Notes, Official Session Note) to prevent confusion.

  • Many practitioners don't fully trust AI to handle sensitive client data, so we also needed to offer no-AI alternatives to help serve that group’s needs.

3

Legal compliance is critical

This meant properly handling consent, ensuring official session notes are properly signed and locked, and addressing concerns around client data duplication (especially for practitioners juggling multiple tools/systems).

design iterations

  • Early iterations of the recording control bar in both video and in-person sessions. The floating bar idea initially received favourable preferences but started to feel too distracting and disruptive for a support tool.

  • Several rounds of revisions for the note editor were required due to changing assumptions around appointment assignment, versioning, and what type of notes ultimately influenced the official clinical note.

  • One particular challenge I grappled with was how important it was to make the manual notes distinct from what the recording generated, but this was deemed unnecessary by the clinical advisor (“…Clinicians take full ownerships of the entire document. The final note becomes a unified clinical document.”) This was a classic example of my initial designer assumptions being corrected by actual user workflow understanding; their real-world perspective helped prevent me from over-engineering the interface :)

  • Early iterations of the recording control bar in both video and in-person sessions. The floating bar idea initially received favourable preferences but started to feel too distracting and disruptive for a support tool.

  • Several rounds of revisions for the note editor were required due to changing assumptions around appointment assignment, versioning, and what type of notes ultimately influenced the official clinical note.

  • One particular challenge I grappled with was how important it was to make the manual notes distinct from what the recording generated, but this was deemed unnecessary by the clinical advisor (“…Clinicians take full ownerships of the entire document. The final note becomes a unified clinical document.”) This was a classic example of my initial designer assumptions being corrected by actual user workflow understanding; their real-world perspective helped prevent me from over-engineering the interface :)

keep reading ツ

𖡼.𖤣𖥧𖡼.𖤣𖥧

keep reading ツ

𖡼.𖤣𖥧𖡼.𖤣𖥧