Scope of the project

What we are building first

SynQ is a dental app for Bangalore built on two centerpieces: a 30-second flow that takes a person from tooth problem to booked clinic to cab, and a universal parsing engine that turns messy healthcare documents into clean, structured records.

Objectives

Win the moment of need

Someone in pain at 11pm should understand what is happening and be booked into nearby care in under 30 seconds, with no dead ends.

Stay outside medical-device regulation

Explain and let the patient choose. Removing app-side triage keeps the MVP clear of CDSCO medical-device territory without losing any patient value.

Build the chassis right

Org + branch from day one and every external integration behind an adapter interface — a vendor SDK is never imported outside its adapter, which makes ChainX a zero-refactor migration.

Own the data moat

Every inquiry records symptom, care preference, clinic, price and outcome — real pricing intelligence, demand maps and honest clinic rankings.

Included modules

MVP means build now. Stubbed means the interface exists and returns mocks. Gated means the module waits behind a compliance milestone.

Symptom Education Library

MVP

~20 authored symptom blocks: what this usually means, what people commonly do, what a dentist does, how to prepare. Template-served, general, never person-specific.

Care Preference Selector

MVP

The patient picks their own care intent — care today / can wait / just info. No label is applied by the software.

Care Router

MVP

Sorts clinics by the patient's choice plus proximity and availability. Same logic a food-delivery app uses. No urgency engine.

Universal Safety Banner

MVP

A fixed, always-visible 108 notice on every screen for every user — general guidance, never a system decision.

Two-Tier Clinic Directory

MVP

Active verified tier (HSR + Koramangala, with pricing) plus passive citywide listings so nobody ever hits “no results”.

Booking Rails

MVP

WhatsApp deeplink (wa.me) as the universal rail, tap-to-call, and a thin SynQ layer for pilot clinics. No Business API, no vendor.

Transport Hand-off

MVP

Ola / Rapido deep link with the clinic address pre-filled. Not an API integration — half a day of work, feels seamless.

Auth & Identity

MVP

Phone + OTP only via Firebase Auth. No email, no social. Patient / dentist / clinic QR codes, built ABHA-ready.

Event Capture & Patient Timeline

MVP

Every column exists from day one. A stateful patient-journey timeline rather than a flat appointment log.

Parsing Engine — Zone A

MVP

Interface, async pipeline and verification gate, exercised on clinic rate cards during onboarding. Zero patient data.

ABDM Service

Stubbed

ABDMService interface with a StubAdapter returning mocks. Eka wired later; the slot is left ready.

Thin SynQ EMR Core

Gated

Patient profile, ABHA link, appointment + visit history, patient↔dentist link. Clinical notes and prescriptions wait behind ABDM HIP certification.

Parsing engine zones

The MVP line: Centerpiece 1 in full, plus Centerpiece 2 at Zone A only. Zones B and C come later, behind compliance gates.

Zone A

In MVP

Non-clinical documents — clinic rate cards and price lists parsed during onboarding. Zero patient data, parse freely now.

Zone B

Post-MVP

“Organise, don't interpret” clinical documents — a patient's old X-rays, prescriptions and history forms. Needs explicit layered per-upload consent, a per-clinic Data Processing Agreement under DPDPA, and flag-for-review before anything lands.

Zone C

Later — strictest gate

Live clinical documentation (“SynQ Scribe”). Per-visit layered consent, raw audio purged fast, and mandatory dentist sign-off.

Functional scope

  • Symptom entry by tap or free text, with plain-English explanation returned from a template library
  • Care-preference selection driving clinic ordering, never clinic filtering by a software-assigned label
  • Clinic cards showing distance, verified badge, response SLA, and price ranges for the active tier
  • One-tap call and one-tap WhatsApp booking with a pre-written, clinic-addressed message
  • One-tap cab with the clinic address pre-filled in Ola or Rapido
  • Off-hours and no-results fallbacks — the app never hits a dead end
  • Consent capture on first launch, stored with timestamp, consent-text version and locale
  • Append-only audit log on every PII access; PII never in a URL; per-user hard-delete job

Business scope

  • Clinics first: hand-pick the first 15–20 clinics in HSR and Koramangala — mid-range, spare capacity, answer their own WhatsApp, willing to share prices
  • The clinic pitch: free patient leads, no upfront cost, pay only for outcomes later
  • Clinic response SLA as a condition of keeping the verified badge
  • Price transparency as the patient-facing wedge — validated with 5 real clinics before it is committed to
  • Zero-budget patient acquisition: founder-authored dental content, local community groups and QR codes at partner clinics
  • Staged launch: private test → invite-only beta → public corridor launch → Month-3 gate
  • Month-3 gate measures 40%+ booking completion, 70%+ clinic response within the promised window, and 800+ corridor downloads
  • True retention signals — return at the next dental moment and family members added — are judged at month 6–9, not on a 30-day reopen
  • Captured data compounds into DentX, ClinicX and ChainX as separate downstream products

Still open — not started until called

  • Booking strategy A / B / C — recommendation is C: deeplink plus a thin SynQ EMR for pilots. Blocked on calling 4–5 pilot clinics.
  • Two-week prototype, or 4–6 weeks launch-ready — this decides how much scope gets cut.
  • LLM explanation layer live, or pre-written templates — recommendation is templates first.
  • Thin EMR core in the MVP — yes or no.

Verified before build

  • A medical-device lawyer confirms that “explain, don't grade” keeps the symptom check outside medical-device rules — the most important legal call before launch.
  • Which EMRs HSR/Koramangala pilot clinics actually use — decides whether the booking option is viable.
  • Clinics will genuinely share prices — validated with 5 real clinics before the transparency wedge is committed to.
  • The citywide passive clinic data source is legally usable.
  • wa.me deeplink opens correctly on Android and iOS with the message pre-filled.
  • Real Ola / Rapido URI schemes — the ones in the specs are illustrative placeholders.
  • PWA versus Android native — push reliability for booking confirmations.

Legal must-haves before public launch

  • Medical disclaimer
  • Privacy policy
  • Terms of service
  • Consent text
  • Named grievance officer

Deliberately out of scope

Each exclusion is a decision, not an omission — most of them exist to keep the MVP outside medical-device and full health-record regulation.

  • Any urgency, severity or triage grading
  • Clinical interpretation of any parsed document
  • Auto-writing anything to a patient's official record
  • WhatsApp Business API, vendor templates or Meta approval
  • Full cab API integration
  • Live ABDM record exchange (interface only, stub adapter)
  • Treatment notes, prescriptions, charting and billing (full HIP load — not in MVP)
  • Zone B and Zone C parsing (patient records and visit notes)
  • Email or social sign-in
  • Pub/Sub event bus and BigQuery warehouse (write directly to Postgres first)