The Science Behind the Platform

CognitiveBotics is grounded in evidence-based behavioural intervention, peer-reviewed research, validation by India's national autism authority, and a granted patent on the underlying methods. This article summarises what the evidence says.

Peer-reviewed clinical evidence

A 12-month longitudinal observational study evaluating the CognitiveBotics platform was published in JMIR Neurotechnology in 2025:

Atturu H, Naraganti S, Rao BR. Effectiveness of Artificial Intelligence–Based Platform in Administering Therapies for Children With Autism Spectrum Disorder: 12-Month Observational Study. JMIR Neurotechnology 2025;4:e70589. doi: 10.2196/70589.

Forty-three children aged 2 to 18 with diagnosed ASD were followed for a year. All received standard therapy; the intervention group also used the CognitiveBotics platform consistently per protocol, while the control group did not maintain consistent platform use. Standardised assessments — CARS (Childhood Autism Rating Scale), the Vineland Social Maturity Scale, the Developmental Screening Test, and the REEL (Receptive Expressive Emergent Language Test) — were administered at baseline (T1) and at the 12-month endpoint (T2).

In the intervention group, every measured domain showed statistically significant improvement (P<.001):

• CARS score reduced from 33.41 (mild-to-moderate range) to 28.34 — moving children below the autism threshold on this scale.
• Social age rose from 22.80 to 35.76 months — a 56.84% gain.
• Social quotient rose from 53.26 to 64.75 — a 21.57% gain.
• Developmental age rose from 30.93 to 45.31 months — a 46.49% gain.
• Developmental quotient rose from 70.94 to 81.33 — a 14.65% gain.
• REEL receptive language improved by 56.22%.
• REEL expressive language improved by 59.93%.

In the control group, while several measures trended in the right direction, none reached statistical significance — underlining that consistent platform use, alongside conventional therapy, is what produced the change.

The authors concluded: "Overall, the platform was an effective supplement in enhancing therapeutic outcomes for children with ASD. This platform holds promise as a valuable tool for augmenting ASD therapies across cognitive, social, and developmental domains."

Government validation — NIEPID

The platform has been validated by the National Institute for the Empowerment of Persons with Intellectual Disabilities (NIEPID), the Government of India's national Centre of Excellence for intellectual disabilities, under the Department of Empowerment of Persons with Disabilities, Ministry of Social Justice and Empowerment. NIEPID's professional therapists and clinicians have reviewed the platform in operational use with children, and the validation letter is on file with CognitiveBotics. NIEPID is also a capacity-building partner for therapy-centre rollouts under the phygital model.

Granted patent

The methods underlying CognitiveBotics are protected by Indian Patent No. 586929, "Methods for Aiding Persons with Autism to Acquire Skills" (Application No. 201941032153, filed 8 August 2019, granted 15 April 2026), with a 20-year term from the date of filing. The patent covers the platform's approach to structuring, delivering, and measuring skill-acquisition activities for children with autism.

What the underlying science is built on

CognitiveBotics combines several well-established behavioural-science traditions:

Applied Behaviour Analysis (ABA). Each Learning Objective is decomposed into discrete targets with a defined mastery criterion (correct first-try response on three consecutive days), mirroring discrete-trial training. The platform records prompted vs. independent responses and surfaces them in the report so therapists can fade prompts at the right pace.

Computer-Assisted Intervention (CAI). Research on CAIs shows they support four key components of effective learning: active engagement, group participation, regular interaction and feedback, and integration with real-life settings. CognitiveBotics is built around all four. CAIs were already showing strong outcomes pre-pandemic; therapist adoption rose from 22% to 65% during COVID, and the field has matured rapidly since.

Video modelling. Where appropriate, LOs use animated demonstration of a target task before asking the child to perform it. Studies show video modelling improves acquisition of conversational, commenting, and pragmatic-language skills in children with autism.

Animated games as a learning vehicle. Engaging animation increases attention span and skill retention for children with autism, and animation-based interventions have been shown to produce significant gains in language and social skills.

Multi-modal AI detection. The platform uses gaze detection, tone/speech detection, and pose detection (PoseNet) so children must physically participate — looking, speaking, moving — not just tap. This produces richer behavioural data and tracks dimensions like eye-contact duration, voice modulation, and posture that pen-and-paper assessment cannot capture in the moment.

Parent-mediated intervention. A consistent finding across the autism literature is that parent involvement is one of the strongest predictors of positive long-term outcomes. CognitiveBotics is designed as a co-viewing product — parents sit with the child, watch parent-training videos, and reinforce techniques throughout the day, rather than handing the child a screen and walking away.

Mastery, not memorisation

A target counts as mastered only when the child gets it right on the first opportunity of the day for three consecutive days. If the child misses a day or answers incorrectly on the first try, the streak resets. This rule prevents single-day luck or short-term memorisation from being mistaken for genuine acquisition. A learning objective is mastered when a defined number of its targets meet that bar (for example, mastering 5 of 6 targets within "Identify Verbs"). The graphical reports track this longitudinally so parents and therapists can see real, durable gains.

What the science does not claim

CognitiveBotics is a supplement to therapy, not a replacement for in-person clinical care. The published study found significant gains specifically when platform use was paired with consistent standard therapy and parent training. The platform is not a diagnostic tool and does not "cure" autism. The intent is to extend the reach and consistency of evidence-based intervention.

Reading the research yourself

The full JMIR paper is open-access at https://neuro.jmir.org/2025/1/e70589. If you have questions about the methodology, the cohort, or how findings apply to your child or your centre, please write to support@cognitivebotics.com.