Helplines, community workers, specialist clinics, recovery programmes, grant-makers and research teams each carry different responsibilities. The common need is continuity with consent: the right context for the right person, without making a vulnerable person repeat their story.
| MH-0418 | PHQ 16 | item 9 = 1 ?? |
| MH-0418 | 16 | called? not sure |
Five places, one person. Nobody can see she's stopped coming.
Many organisations combine two or three. Start with the closest match, then borrow only the parts that solve a real programme problem.
Community workers identify needs, provide structured support and escalate when required.
Clinical teams assess, treat and review people over multiple episodes of care.
People reach trained responders by phone, chat or scheduled remote sessions.
Support combines counselling with family work, housing, livelihoods and community participation.
A funder supports multiple mental-health organisations and needs portfolio learning.
Teams run studies, pilots or service-improvement work under formal protocols.
Not sure which one you are? Bring one tracker and describe a normal week. The model usually becomes clear quickly.
These come up in almost every first conversation we have with mental health teams.
"Half our positive screens never got a call back."
Camp scores sit in one sheet and intakes in another, so nobody sees who's still waiting.
"I carry my high-risk clients in my head."
Risk calls happen on the phone at night, and there's no record the supervisor saw them.
"Half her distress was a pension that never came. We had nowhere to record it."
Benefit applications live in a volunteer's notebook, so nobody knows which ones stalled at the taluka office or what happened to the person waiting.
"We can't say who was still well at six months."
Follow-up calls happen, but the people we couldn't reach quietly drop out of the numbers.
We haven't published a mental health case study yet. Community mental health programmes need two things we've built: restricted case records for sensitive disclosures, and scheme applications tracked until the benefit arrives.
Students, trainers and sessions in one system, with location-stamped trainer attendance, a restricted case module for abuse disclosures and donor reports.
Read the case study →Tapasya Foundation5,000+families matched to schemesA master list of government schemes and their eligibility rules matches each family to the right ones, and each application is tracked until the benefit arrives. Live in four months.
Read the case study →Recognise your programme? Tell us where the work gets stuck and we'll suggest a useful first version.
This is one operating model, not a template for every mental health organisation. Use it to see how everyday work, follow-up and management decisions can connect.
Stage 01 · Screen with care
People say tension, a heavy heart, thinking too much. Nobody says depression. A screen only helps if every positive score, and every positive item 9, reaches someone who can respond that week.
| ID | Setting | Score | Status |
|---|---|---|---|
| MH-0427 | Home visit | PHQ 14 · item 9 | Call pending |
| MH-0431 | PHC OPD | PHQ 12 | Intake due |
| MH-0418 | Health camp, Sinnar | PHQ 16 · item 9 | Supervisor called |
| MH-0409 | School referral | GAD 11 | Intake booked |
Stage 02 · Intake & triage
The first one-on-one decides whether someone comes back. It's also where risk gets assessed properly and a safety plan is made together with the person.
| ID | Risk | Care tier | Status |
|---|---|---|---|
| MH-0427 | High | Specialist | Psychiatrist today |
| MH-0431 | Low | IPT-G group | Triage pending |
| MH-0418 | Moderate | IPT-G group | Safety plan done |
| MH-0409 | Low | Individual HAP | Placed |
Stage 03 · Sessions that work
Brief structured therapies work when they're delivered as designed. In the field, sessions get shorter, content drifts and groups thin out by week four.
| Group | Counsellor | Attendance | Action |
|---|---|---|---|
| Sinnar IPT-G 3 | Anjali Pawar | 62% | Check-in calls |
| Igatpuri IPT-G 1 | Rekha Shinde | 74% | Fidelity review |
| Nashik HAP | Farida Khan | 81% | 2 not improving |
| Trimbak IPT-G 2 | Sunita More | 90% | On track |
Stage 04 · Stay well
End-of-therapy scores look good. What matters is whether people are still well three to six months later, back at work, managing their household, and getting the pension or disability support they applied for.
| ID | End score | 6 months | Status |
|---|---|---|---|
| MH-0352 | PHQ 5 | Not reached | 3 attempts |
| MH-0366 | PHQ 4 | PHQ 13 | Booster offered |
| MH-0341 | PHQ 3 | PHQ 2 | Remission |
| MH-0418 | PHQ 6 | Due Mar | Scheduled |
Choose the parts that match your duty of care. A helpline needs a different record from a clinic, and a grant-maker should not receive identifiable clinical notes.
Set purpose, access, sharing and retention rules before collecting sensitive information.
Give teams clear risk prompts, escalation paths, named responsibility and closure.
Connect screening, low-intensity support, specialist referral and return to community care.
Track training, observed practice, case consultation and supervisor actions.
Manage shifts, interactions, anonymity, follow-up and referral directories.
Add social-inclusion plans, portfolio milestones or protocol controls only where needed.
Most AI projects stall because the records underneath are messy. So we work in a fixed order.
Every field has to earn its place by helping someone decide something. If nobody uses it, we don't collect it.
A system people avoid is worse than a spreadsheet. We design around shared phones, patchy signal and WhatsApp.
On clean, connected records, AI can read paperwork, draft reports and flag who needs attention. A person always checks.
Where you start depends on where you are. If your records are already in good shape, we can move to AI quickly. If they're in notebooks, we start there, and we don't start from zero.
Offline forms, local languages, links to government IDs and AI features are set up to suit each programme's needs and consent rules, so not every build includes all of them.
We picked the jobs that save mental health teams the most hours. Each starts with a problem we hear often, and each keeps a person in charge of the decision. How we think about AI →
“We screen people who can't read, and a volunteer reading questions aloud changes how people answer.”
Your team still decides: human recordings only, and a supervisor spot-checks the audio each quarter.
“Nobody in the village says depression, and our awareness films still use words people don't use.”
Your team still decides: local volunteers and community members review every script first.
“We find out someone stopped coming only when the volunteer mentions it at the monthly meeting.”
Your team still decides: the supervisor goes through the list in weekly supervision.
“Our supervisors can't listen to every session, so feedback reaches volunteers weeks late.”
Your team still decides: the supervisor listens to the flagged parts and owns the rating.
“Our volunteers spend days filling the same details into pension and disability forms, and half come back for missing documents.”
Your team still decides: the volunteer checks every field with the person before anything is submitted.
“Writing funder reports means reading through case notes, and in a small village one detail can identify someone.”
Your team still decides: the programme lead removes anything that could identify someone in a small community.
Mental health data is some of the most private there is. These lines don't move, whoever's asking.
She sees when someone she screened was contacted and started care.
A clear triage decision, supervisor backing on every risk call, and each client's score trend.
Risk changes, people not improving and fidelity by skill, ready for supervision.
Open places by language and location, and waiting times, with Client IDs only.
Response and remission by cohort, with people lost to follow-up shown as their own number.
Six-month follow-up is where most programmes lose people, and the people you can't reach are rarely the ones doing well. We report remission and reach, side by side, every time.
It's the number funders trust, because it's the hardest one to show.Most systems launch in three to four months. The timeline moves with the number of programmes and roles, how messy the old data is, and how many tools need connecting.
Tell us where the work gets stuck today. Spreadsheets, notebooks, WhatsApp groups, all of it.
You show us how the programme runs. We'll tell you straight if you need a new system, a better workflow or nothing at all.
We map how your programme works, talk to the people who'll use it, review the tools you use today and hand you a System Blueprint, a clear plan of what we'll build.
Working screens every 2 to 4 weeks, tested by your field team in your own terms. We clean up your old spreadsheets and move them across.
Field training and launch, then 30 to 60 days of close support. After that, the system keeps changing as your programmes do.
We connect to IVRS and WhatsApp, Tele-MANAS referral records, your research exports (Stata, R or SPSS) and partner organisations' own sheets, rather than replacing what already works. Zoho licences sit in your organisation's name, and everything exports whenever you want it.
Research-led programmes. If the programme sits inside a study, we work with your PI and ethics requirements from the start: consent versions, trial arms and de-identified exports.
The questions your finance head and IT lead will ask, answered before the call.
Zoho licences are bought in your organisation's name and hosted in the Zoho data centre region you choose, including India. We work inside your account.
Every record exports to Excel or CSV whenever you want, including the day you stop working with us. You're not locked in to us for support either.
Each person sees only the records their role needs. Built to respect confidentiality under the Mental Healthcare Act 2017. Names sit apart from scores and session notes.
Edits are logged with who made them and when, so an audit or a funder query doesn't start from memory.
Consent is recorded for each purpose, and we agree up front how long data is kept and when it's deleted. ZolaShield, our DPDP platform, shows where you stand.
AI features run on your records under your account's controls. Beneficiary records never go into public AI tools.
We’ll map what your team does now, identify the smallest useful first version and show where follow-up, learning and reporting can share the same information, without forcing your programme into a generic template.
If budget is tight, ask about the Built for Impact Programme, which gives eligible nonprofits free Zoho credits.
The same approach, shaped to each kind of work.
