In-school social skills groups for autistic teens, with parent coaching
TL;DR Socialcircle runs a 14-week social skills programme for autistic students aged 11–17 inside mainstream private schools. Small groups meet weekly with a trained facilitator and classmates join as practice partners, while parents get their own online session. Schools pay per group. It swaps the seed's headsets for a format that has trial evidence behind it.
Autistic teenagers in Indian mainstream schools usually fall behind socially long before their marks slip. In a Bengaluru-area study of 43 autistic teenagers and young adults, 31.7% had dropped out of school because of behavioural and academic difficulties, and families described isolation and burnout[1]. The government counted about 22.7 lakh children with special needs in schools by 2022[2], and autism appears in India at rates similar to the rest of the world[3].
So what do schools do about it? Mostly nothing structured. They admit these students but rarely have a way to teach conversation, joining a group or handling teasing.
The help that does exist is one-to-one therapy outside school, far away from the classmates the child actually has to get along with, which in my opinion misses the point.
What it does, in regulatory terms: structured social-skills teaching in school, not therapy or diagnosis. Group leaders are special educators or psychologists on the Rehabilitation Council of India's register, since only registered professionals may practise[4]. Socialcircle makes no claim to treat autism.
Here's how it works. Six to eight students meet once a week for 14 weeks during school hours. Each session teaches one skill, like entering a conversation or handling a disagreement, and then the group practises it in role play with two trained classmates. Parents join a weekly online session to rehearse the same skill at home. Facilitators keep a short skill checklist each week, and the school gets an end-of-term report.
The seed pitched augmented reality with AI feedback, and I'd drop it. Stanford's wearable trial produced a modest gain that faded six weeks after it ended, and families used the glasses for about half the planned time[5]. A 2026 virtual reality study saw skills improve inside the headset but not in the classroom two weeks later[6]. Plain old group training shows a moderate effect across 25 randomised trials[7].
The student benefits but the school pays per group, out of its inclusion budget, and it can pass some of the cost to parents as an activity fee. I couldn't find a public price for social skills groups in India. The nearest benchmark is individual teletherapy at ₹800 for a 45-minute session from an Indian provider[8]. Start with a per-group fee per term and test it in the pilot.
Two meta-analyses published in 2025 put the effect of social skills training in the small-to-moderate range, which is honest enough to sell on - no miracle claims needed[7][9]. The 2026 virtual reality result gives schools a reason to prefer in-person practice[6]. And the DPDP Rules, notified in November 2025, set out how parental consent works for children's data, which matters when home practice involves video[10].
There's no public figure for school-based social skills programmes in India. The closest number is the 22.7 lakh children with special needs in schools[2], of whom autistic teenagers in fee-paying private schools are a small slice. Count the schools you can reach in one city before you model anything bigger.
A Delhi autism non-profit trains special educators and runs education programmes[11]. A Mumbai child-development non-profit mentors therapists online through an autism tele-mentoring programme[12]. A UCLA-developed curriculum is the best-known manualised social skills programme and certifies providers worldwide[13]. One US startup sells virtual reality practice and has an FDA breakthrough designation, not a clearance[14]. Nobody I found in India sells a term-long group programme to schools. That's the gap.
| Difficulty | Moderate (36/100) |
| Starting capital | ₹70 L–1.2 Cr |
| Time to launch | 4–6 months |
There's very little to build. The hard part is adapting a curriculum to Indian classrooms and languages, then training facilitators so session 9 in Pune looks like session 9 in Kochi. Schools also buy slowly, usually once a year, and want peer references first.
The business needs no special licence, but the people who run groups must be registered, and the data is sensitive.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
AI should never judge a child's social skills, coach them directly or make claims about autism. The evidence is for trained people running groups, and the 2026 virtual reality study is a warning that practice in a tool may not carry over[6].
Want to work out the AI layer for Socialcircle? Talk to Yash →
I like this as a services business for a psychologist or special educator who has run groups and can sell to school principals. If you want to build hardware then this isn't it. Prove one thing first - two schools pay for a second term after seeing parent and teacher ratings from the first.
The free Socialcircle validation kit has the line-by-line capital breakdown, the effort scoring, the registration and child-data checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for parents and school leaders, and 5 assumptions to test. Download the kit →
Thinking about building Socialcircle or something close to it? The make-or-break bit is showing that skills carry over from the group into the rest of the school day, and that's the kind of thing I help founders with. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.