UX Design | Product Design | User Research
AIDOK is a light-weight and durable earpiece which utilizes AI and speech-to-text-to-speech tehnology to revolutionize emergency care. It provides paramedics the possibility to document, communicate, and to use translation/interpretation feature in real-time and at the same time freeing their hands for patient care.
This was a really extensive project, so not everything can be presented here. Our client Solita Oy was very happy and pleased with everything we did and the outcome. Learn more about the project in detail from our blog on Notion
- Client: Solita Oy
- Target group: Paramedics
- Team: 6 members (with me being the only Finnish speaking)
- Course: HTI.230 Human-Centered Product Development (Tampere University)
- Explore user-centric, futuristic ideas for leveraging technology to support paramedics in their tasks
- Innovate a hands-free solution for ensuring paramedics have all necessary information during patient care and to improve efficiency
- Understand paramedics’ duties and identify when hands-free UX is required
- Report on how hands-free UX functions and the interaction techniques used to control applications and complete tasks
- A dialog-based concept (concept video)
- Storyboards
- User persona
- Product design
- Project presentations
- Project blog (detailed information of everything we did and found out)
I had a key role throughout the project as the only Finnish speaker on the team.
- Conducted extensive desk research
- Recruited paramedics for interviews, prepared all things for the interviews, and participated in all sessions (as an interviewer, or observer & interpretator)
- Created the user persona and two storyboards (digitally drawn with a Wacom pen tablet and Canva)
- Designed all the questionnaires
- Contributed to data analysis collaboratively
- Handled all Finnish-English-Finnish translations and interpretations
- Assisted in concept ideation (excluding product design images)
- Participated in role-playing for the concept video (as a paramedic)
- Prepared and delivered presentations
- Contributed actively to the project’s blog
A lot of different kinds of methods were used during the project. I'm not going to go into details regarding each of those, but if you're interested in learning more about the methods and the project in general, an extensive blog was written during the course. Take a look at our blog in Notion.
We read over 25 articles and searched for possible current and competing devices, learned about the trends, AI, LLM, generative AI, etc. Here's a list of some of the things we took a look at:
- Virve 2.0
- AI & Generative AI
- LLM
- Lifecare
- Merlot Medi
- VuzixM400 Smart Glasses
- GlassE Smart Glasses
- Speech Based Assistants
- SnapCap System
- Using ChatGPT in Healthcare
- N=5
- Semi-structured interviews + background questionnaire
- Thematic content analysis
Paramedics’ Pain Points
- Documentation
- Cognitive load
- Accessing data
- Sharing information
- Uncertainty
Other Important Findings
- Different wellbeing services counties often have different kinds of systems and devices, and therefore a bit different ways of doing things
- Audio-based user-interfaces may be limited in noisy environments
- The system must be trustworthy, and privacy issues need to be considered
I created this with Canva.
Our original plan for the product was to create a pair of smart glasses (“NursEyes”) with built-in generative AI capabilities in addition to communication and translation features. However, the evaluation of this 1st MVP with paramedics gave us important information that led to a major revision in our original idea.
- N=6
- Background questionnaire for new participants
- Semi-stuctured interviews
- Post-questionnaire
- Form factor (glasses) was disliked
- Must be reliable and accurate
- Live translation/interpretation would be really useful
- Wouldn’t like lots of audio-based information
- Mixed feelings about visualizing patient data “on the patient”: some might like it a lot, but they can already read information from the screen(s)
- Should support collaboration
- Not one solutions fits all
- Transcription and GenAI to record and summarize speech-to-text recorded notes and save captured images.
- Communication via different channels, and streaming live video to, for example, share patient’s condition to a doctor.
- Easing communication with patients, eyewitnesses, etc. without a common language.
I created this with a Wacom pen tablet and Canva.
I created this with a Wacom pen tablet and Canva.
We created a role-played concept video to better demonstrate our concept idea and the interaction. The concept video, as well as the storyboards and product design, was presented also to the interviewed paramedics during the evaluation of our 2nd MVP (AIDOK).
Watch the concept video on YouTube
The MVP of AIDOK was evaluated with semi-structured interviews and a post-questionnaire. Additionally, we had a background questionnaire for new participants. The data was analysed using an affinity diagram.
- N=5 (60 % females, 40 % males)
- 32-48 years
- 4 wellbeing services counties
- The concept and all of its features were liked a lot, and so were the storyboards and the video
- The translation/interpretation feature needs more iteration
- Focus should be on reliability, accuracy and safety & privacy issues
- Integration with their other main devices/systems is a must
Please, have also a look at our final presentation included in this repository. It contains a more detailed description of the findings regarding, for example, the features, concerns, other topics and feedback, and the results from the post-questionnaire.
Do you want to know more about the project? This was a really extensive project, so everything cannot be presented here. Our client Solita Oy was very happy and pleased with everything we did and the outcome.