An Indian start-up's AI platform builds individualised speech, behavioural and social-emotional learning plans for neurodivergent children; a 12-month, company-funded Hyderabad study reported large gains for 35 children but used only 5 untreated controls.
35 children
Children followed on the platform (12 months)
5 children
Untreated comparison group size
15 %
Reduction in CARS autism-severity scores
46-60 %
Gains in social age, developmental age and receptive/expressive language
Details
Maturity
Scaling
Promoter
CognitiveBotics Technologies Private Limited, with Rainbow Children's Hospital and CARE Hospitals
Period
2023–2025
Keywords
special education, assistive technology, autism therapy, ADHD and neurodivergent support
Context
Families of neurodivergent children in India, including those with autism, ADHD and Down syndrome, often face long waits and high costs for specialist speech, occupational and behavioural therapy, and CognitiveBotics, a Hyderabad start-up founded by parents of a child with autism, built an AI-based eLearning platform to extend and monitor therapy between clinical sessions for children aged 2 to 18.
Objectives
The platform was designed to use AI to generate individualised, gamified learning plans across speech, occupational, behavioural and social-emotional domains, letting therapists remotely track each child's progress and adjust the plan, developed with and used alongside clinical teams at Rainbow Children's Hospital and CARE Hospitals in Hyderabad.
Activities
A 12-month observational study published in JMIR Neurotechnology (April 2025) followed 35 children using the platform alongside continuous therapy and compared their outcomes against a 5-child untreated comparison group.
Results
The study found statistically significant improvements versus the untreated comparison group, including a 15% reduction in CARS autism-severity scores and 46-60% gains in social age, developmental age and receptive/expressive language.
Conclusions
The study was funded and sponsored by CognitiveBotics itself, with its lead author sitting on the company's advisory board and its corresponding author a company employee, a conflict of interest disclosed in the paper; the tiny control group and lack of independent replication mean these results should be read as promising but unproven rather than confirmed.
Implementation
Indicative cost
Medium (€50k–€500k) — No public pricing disclosed; the platform requires ongoing clinical partnership with hospital-based therapists rather than a one-off purchase.
Time to results
Medium (1–3 years) — Platform developed 2023; 12-month evaluation study published April 2025.
Staffing & skills
CognitiveBotics platform team, Clinical therapists at Rainbow Children's Hospital and CARE Hospitals
Conditions for success
Platform used alongside, not instead of, continuous clinical therapy
Remote therapist tracking allowing individualised plan adjustment
Common failure modes
Efficacy study funded and sponsored by the vendor itself, with a disclosed conflict of interest through its lead and corresponding authors
Comparison group of only 5 untreated children with no independent replication
Where it fits
Governance type
start-up platform co-developed with private hospital clinical teams
Scale
35 children across partner hospitals in one city
Income level
lower-middle-income
Replication kit
Reusable artefacts from this practice — as published by their sources.
An EdUHK three-arm RCT found a NAO robot delivering 12 weekly social-skills sessions produced significantly greater communication and social-interaction gains …
A Hangzhou Dianzi University-led team built ReadCraft, letting children with dyslexia co-create illustrated stories with AI before practising reading them. …