Awareness without action
Mental health awareness in urban India has risen sharply. "Therapist near me" searches are up 3× since 2021. Platforms report record inquiry volumes. The conversation is everywhere.
Yet the treatment gap for common mental disorders in urban India remains 70–80% — even among educated, working adults in metropolitan cities.
This study asked a narrow question: what is happening in the decision space between "I think I need help" and "I am sitting in a therapist's office"? Who are the people in that gap, and what would actually move them?
Decision architectures, not demographic profiles
We built 8 deep behavioral personas representing distinct delay archetypes — each grounded in NMHS data, Lancet India research, iCall annual reports, and Reddit India forum analysis. Each persona was placed at their decision moment and probed through 5 research questions and 4 intervention conditions.
Total simulation interactions: 40+. Recruitment cost: ₹0. Time to results: under 4 hours.
Every persona has: a specific social audience whose judgment shapes decisions, a named barrier mechanism, a prior exposure history, and a precise trigger condition. The result is simulation outputs that tell you not just what people do — but exactly why, and what specifically changes it.
8 personas. 7 distinct barrier types.
Each persona was built to represent a specific decision architecture — not a demographic segment. The barriers are named precisely, not categorically.
Bengaluru
Mumbai
Delhi
Hyderabad
Pune
Delhi
Hyderabad
Mumbai
Six things the simulation revealed
Stigma campaigns are solving the wrong problem for 75% of the audience
Awareness campaigns scored 0–2/10 across all 8 personas. Stigma is the primary barrier for approximately 1 in 8 delayed individuals. The majority face identity threat, process friction, role obligation, category misclassification, or inertia.
The most common barrier in the cohort is not "I'm ashamed to go." It is: "I haven't placed myself in the category of someone who needs it." These two states require completely different interventions.
Free therapy does not equal used therapy
For 6 of 8 personas, cost removal was the weakest or second-weakest intervention. Vikram — a senior consultant earning ₹50L/year — scored price reduction at 0/10. Rohan at 1/10. Even Arjun, who is genuinely financially constrained, scored it 3/10 — because without clinical naming, free access doesn't convert.
Cost is necessary for the economically constrained sub-population. It is not sufficient for anyone. The field's focus on affordability, while warranted for equity reasons, is not the primary lever for urban, educated, working Indians.
The invisible threshold: benchmark degradation
The "not sick enough" mechanism operates through a pattern we named benchmark degradation: the personal standard for "fine" shifts downward over time without the person noticing. Vikram has been in burnout for 14 years. His current baseline is his year-10 burnout state — not a healthy state. He is comparing himself to a degraded reference point.
This is the most dangerous sub-type: high-functioning, self-assessed as managing, actually in significant decline. Standard interventions don't reach them because they don't believe they're in the relevant category. The only entry point is data about the blind spot itself.
The process friction paradox
Zara and Divya have the highest mental health literacy in the cohort, the most positive attitude toward therapy, and zero stigma. They are also among the most delayed. Their barrier is entirely operational: they don't know how to find a therapist, evaluate a profile, or know if their problem is serious enough.
The optimal intervention is the opposite of marketplace logic: one specific person, one price, one available slot, delivered at the moment of peak intent. When Zara's friend sent her a single name, price, and contact detail, she booked within 72 hours after 6 weeks of inaction. The barrier was never attitudinal.
The proxy pathway is almost entirely absent from service design
Meena does not respond to any standard mental health intervention. She responds to: her counsellor asking "How are you doing?" and framing support as "helps you help your son." The "come in for yourself" entry point fails for an entire generation of women whose identity is organized around being the stable provider for others.
She accepted a referral for herself in the same helpline call where she was seeking help for her son. The proxy frame was the only architecture that didn't require self-prioritization. She does not experience this as seeking help. She experiences it as becoming more effective at helping him.
Intent decays within 48 hours without a simultaneous action pathway
Peer normalization is the highest-converting intervention in the cohort — but only when paired with an immediate booking pathway. Divya's intent after reading a specific, relatable LinkedIn post was 8/10. Without a booking link in the same post: no booking by day 3. With a booking link at the bottom: booked in 4 minutes.
The design implication is precise: the booking pathway must appear at the moment of peak intent — not on a separate page, not after account creation, not one email later. In the same scroll. This is a product design requirement, not a messaging requirement.
Four designs. Each one administered as a live re-simulation.
Each intervention was tested by placing the persona directly inside the scenario — a message they received, a conversation they had, a post they saw. The persona reacted in real time. Intent-to-seek was scored before and after each exposure.
Mean intent: 2.6 → 8.5
All 8 personas re-simulated with their targeted intervention. No persona remained below 7/10 post-intervention. 5 of 8 reached definite booking; 3 reached likely booking within one week.
| Persona | Barrier type | Baseline | Post-intervention | Outcome |
|---|---|---|---|---|
| Anika, 28 | Information management | 3/10 | 8/10 | Booking likely (72hrs) |
| Rohan, 34 | Identity threat | 2/10 | 8/10 | Booking likely (1 week) |
| Zara, 23 | Process friction | 4/10 | 9/10 | ▲ Booked |
| Arjun, 26 | Clinical naming gap | 2/10 | 8/10 | Conditional on cost |
| Sunita, 38 | Role obligation | 3/10 | 10/10 | ▲ Booked immediately |
| Vikram, 41 | Category misclassification | 1/10 | 7/10 | Booking likely (1 week) |
| Meena, 44 | Frame invisibility | 1/10 | 9/10 | ▲ Booked same call |
| Divya, 31 | Inertia gradient | 5/10 | 9/10 | ▲ Booked (4 minutes) |
Why deep personas produce non-obvious findings
The gap between generic AI persona output and Simulatte output is the gap between a demographic description and a decision architecture. The difference matters because the wrong mechanism leads to the wrong intervention.
| Generic AI output | Simulatte persona |
|---|---|
| "28F, urban, open to therapy, worried about stigma and family judgment" | Anika delays because of a specific information cascade risk in her arranged marriage timeline. Her personal stigma score is 0.22 — very low. Her social consequence score is 0.85. Remove the information risk and she books this week. Stigma reduction campaigns do nothing for her. |
| "Male professional, stressed, avoids help-seeking due to masculinity norms" | Rohan's barrier is one unconscious equation built around a single event at age 14. His father's breakdown was never named by the family. Therapy = becoming that man. The intervention is a performance reframe from a male peer — not a stigma campaign and not masculinity deconstruction. |
| "First-gen engineer, cost-sensitive, may benefit from subsidized access" | Arjun is offered free therapy and still hesitates — because he doesn't know social anxiety disorder exists as a clinical category. He believes therapy is for "emotional problems from childhood." Clinical naming must precede subsidy. Cost removal alone is insufficient. |
This study ran in under 4 hours. Zero recruitment. Zero IRB.
8 deep personas. 40+ simulation runs. Baseline findings, intervention design, re-simulation, and case study output — all synthetic, all traceable, all specific to the Indian behavioral context.
If you're researching health behavior, EAP utilization, mental wellness product uptake, or any question where "why don't people act even when they know they should?" is the core problem — we can run a custom study for your question.