KidsAbility has teamed up with the University of Waterloo to tackle a long-standing challenge in pediatric speech-language pathology: the documentation burden. While AI tools like AutoScribe have made strides in occupational and physiotherapy settings, they’ve struggled to meet the unique demands of speech-language pathologists (SLPs)—especially in the dynamic, play-based world of pediatric care.
Speech Language Pathologists at KidsAbility often spend 20+ minutes per session on clinical notes, limiting time for direct client care. Existing AI scribes fall short in capturing the nuances of children’s speech and the interactive nature of therapy sessions. To address these barriers, Kidsability partnered with Tia Tuinstra, a systems design engineering researcher from the University of Waterloo, supervised by Dr. Bryan Tripp. Together with KidsAbility clinicians, Tia developed a custom AI pipeline tailored to pediatric SLPs.
To address these barriers, Kidsability partnered with Tia Tuinstra, a systems design engineering researcher from the University of Waterloo, supervised by Dr. Bryan Tripp. Together with KidsAbility clinicians, Tia developed a custom AI pipeline tailored to pediatric SLPs. Tia Tuinstra, a systems design engineering researcher from the University of Waterloo, supervised by Dr. Bryan Tripp. Together with KidsAbility clinicians, Tia developed a custom AI pipeline tailored to pediatric SLPs. Key innovations include:
- A language model trained on SLP-specific terminology and SOAP note templates
- Advanced transcription that distinguishes speakers and captures speech sounds
- Phoneme-level analysis to assess children’s articulation accuracy
Early Wins & Lessons Learned:
Initial testing showed promising results:
- Tailored templates improved note accuracy and organization
- Structured sessions yielded better AI performance than play-based ones
- Speech sample analysis showed potential, though not yet at clinician-level precision
This collaboration blends cutting-edge engineering with clinical expertise to build AI tools that truly support pediatric SLPs—freeing up time for what matters most: helping children and families thrive.
What’s Next:
The partnership is pushing forward with:
- Enhanced prompts and reasoning for better note generation
- Improved transcription quality and speaker differentiation
- Broader applications like referrals and progress summaries
- Integration with health record systems for seamless workflows
