06 — India-First Route Map

AutoSpeak B2B · v1.0 · 2026-06-02 The global vision (00) is the destination; India is the launchpad. This document explains why AutoSpeak starts in India, what makes the Indian market distinct for voice AI, and how the product is shaped to fit it. Pairs with the vertical menu in 07.

⚠️ Directional overview, not legal or tax advice. India telecom and data rules are actively evolving (Telecom Act 2023 rules, DPDP rules) — validate current specifics with an Indian telecom/corporate lawyer and your telephony provider before launch.

1. Why India first (the strategic logic)

Reason	Detail
*Voice is the* interface for Bharat**	22 official languages and hundreds of millions of users more comfortable speaking a regional language than typing or using English. Voice AI in vernacular is a genuinely India-shaped opportunity that English-first tools handle poorly.
Massive, phone-first SMB market	Clinics, coaching centers, real estate, restaurants, dealerships, salons — high call volume, a strong missed-call culture, and a large cash-on-delivery economy.
A proving ground for global	Win India for traction and case studies, then expand to higher-value markets (US/GCC) on the same platform.

The catch (we plan for it): Indian SMB willingness-to-pay is lower and sales cycles are price-sensitive. AutoSpeak responds with affordable INR pricing, a WhatsApp-native experience, and an efficient architecture rather than premium per-minute pricing.

2. India telephony reality — the single biggest difference from the global plan

India's telecom regulation (TRAI/DoT) restricts how cloud and VoIP services connect to the public phone network (PSTN). For domestic India voice at scale, AutoSpeak leads with a licensed Indian cloud-telephony provider rather than assuming a single global carrier.

2.1 Indian cloud-telephony landscape

India has a mature ecosystem of licensed cloud-telephony providers offering virtual numbers, IVR, programmable call flows, and DLT handling. AutoSpeak's design supports integrating with established India-native providers for domestic calling while retaining a global carrier for international and outbound traffic.

Architecture impact: AutoSpeak's TelephonyProvider abstraction (02 §3.1) earns its keep here — a domestic India adapter sits behind the same interface as the global carrier adapter. A key technical requirement is real-time, bidirectional media streaming between the carrier and AutoSpeak's media worker; where native streaming isn't available, SIP trunking into AutoSpeak's own media server is the fallback path.

2.2 Mandatory: DLT registration (TRAI TCCCPR)

Commercial communications in India must go through DLT (Distributed Ledger Technology) registration operated by the telcos:

Register the business entity on a DLT platform.
Register headers/sender IDs and consent templates.
Map use cases (transactional vs. service vs. promotional) — promotional is the most restricted.
Indian telephony providers typically help with DLT onboarding.

2.3 Calling rules

DND / NCPR scrubbing applies to promotional calls; transactional, service, and consented calls have more latitude — AutoSpeak leads with inbound and consented/transactional flows.
Honor opt-outs, respect calling hours, and disclose AI (best practice now, likely required later).
The Telecommunications Act 2023 modernizes the regime and reinforces consent and anti-spam penalties — we track the rules as they finalize.

3. The Indian-language AI stack — the differentiator

English-only voice AI loses most of India. Multilingual and code-mixed ("Hinglish") handling is both a requirement and a strength. AutoSpeak blends global models for English with Indic specialists for vernacular.

3.1 Languages to target (phased)

Phase 1: English + Hindi + 1 regional language tied to the pilot city (e.g., Tamil in Chennai, Telugu in Hyderabad, Kannada in Bengaluru, Marathi in Pune/Mumbai).
Phase 2: add Bengali, Gujarati, Malayalam, Punjabi, Odia per demand.
Handle code-mixing (Hinglish and similar) — common and technically hard; test explicitly.

3.2 Indic voice and language model approach

India has a rich and growing ecosystem of Indic speech and language models — including strong open research models and capable commercial providers — spanning:

Indic ASR (speech-to-text) with telephony-grade vernacular support.
Indic LLMs for dialogue, including models that can be tuned on Indic and vertical-specific data.
Indic TTS (text-to-speech) for natural-sounding regional voices, with premium multilingual voices available where higher fidelity is needed.
Public language resources and datasets (such as national language-technology initiatives) useful for broad coverage and tuning.

This maturing open and commercial Indic ecosystem is a major reason India is an attractive starting point: the language coverage AutoSpeak needs is increasingly available and improving quickly.

How this maps to the architecture (overlay on the phased model strategy in 02 §4):

Early phases (orchestrate): integrate capable Indic ASR/TTS and a strong dialogue LLM, with premium voices for English where it matters.
Later phases (optimize): as volume grows, move toward tuned and self-hosted Indic models for greater control over quality, latency, and cost.

4. Payments in India (different from the global plan)

Billing the business: support India-native payment gateways and methods — UPI, cards, netbanking, wallets — with monthly and annual INR plans plus usage-based billing. Use a provider that handles RBI e-mandate/auto-debit rules for recurring payments.
In-call payments (e.g., an AutoSpeak Stay deposit or an order payment): the clean India path is a UPI payment link / collect request / QR delivered over SMS or WhatsApp rather than capturing card details on the call. This avoids handling raw card data and aligns with RBI card-on-file tokenization rules — a simpler, safer story than the global card flow.
COD-related flows (see 07): "confirm your cash-on-delivery order / pay now via UPI to confirm" — a high-value, India-specific use case enabled by UPI links.

5. WhatsApp is non-negotiable in India

India runs on WhatsApp. The product is designed as voice + WhatsApp, not voice alone:

Send confirmations, payment links (UPI), summaries, reminders, and follow-ups via the WhatsApp Business API.
Support flows like "missed call → WhatsApp follow-up," "call summary to WhatsApp," and "book via voice, confirm via WhatsApp."
WhatsApp is treated as a first-class channel in the runtime — an output adapter alongside SMS and email.

6. India entity, tax & compliance setup

Private Limited Company (incorporated via MCA), with founder agreements and IP assignment.
DPIIT Startup India registration (tax benefits, easier compliance, government schemes).
GST registration and INR invoicing.
DPDP Act 2023 readiness: consent and notice, data-principal rights, Data Fiduciary duties, grievance officer, and a breach process; watch for additional duties if classified a Significant Data Fiduciary. (See 03.)
DLT registration (§2.2) and telephony-provider compliance.
Core documents: Terms of Service, Privacy Policy (DPDP-aligned), Data Processing Agreement, and an Acceptable-Use Policy (no spam/scam calling — protects number reputation).
Appropriate insurance (cyber / E&O) given the handling of voice and personal data.
Global structure later: when raising global capital, a common path is to add a US parent company with the India entity as a subsidiary.

7. India go-to-market

7.1 Beachhead segments (high call volume + clear ROI + reachable)

Lead with one or two, then expand using 07:

Healthcare — clinics, diagnostic labs, dental, multi-specialty (appointment booking, reminders, report queries). High volume, clear no-show ROI.
Real estate — inbound lead qualification and site-visit scheduling (developers and brokers spend heavily on lead gen; missed calls mean wasted ad spend).
Education — coaching centers, edtech, schools, colleges (admission inquiries, fee and parent queries, seasonal spikes).
Restaurants / cloud kitchens — reservations and takeaway/order capture.
Automotive — dealership service booking and test-drive scheduling.
E-commerce / D2C — COD order confirmation and delivery coordination (reduces fake-COD losses — a clear India ROI story).

7.2 Motion & pricing

Direct SMB sales in dense clusters (a city's clinics, a developer's projects) → reference density → word of mouth.
Partnerships: integrate with vertical SaaS (clinic management software, real estate CRMs, restaurant POS) and telephony resellers; co-sell.
Pricing: affordable monthly INR plans with included minutes/calls and overage, or per-call pricing for transactional flows (COD confirmations, reminders). Keep it simple and accessible for SMBs. Bundle WhatsApp.
Demo line: a "call this number to talk to our AI in Hindi or English" demo converts skeptics quickly.

8. India phased journey (overlay on the roadmap in 01)

 INDIA — Foundation
   incorporate · DLT · domestic telephony · payments · Indic stack · DPDP basics · WhatsApp API
 INDIA — Pilot (1–2 verticals, 1–2 cities)
   English + Hindi + 1 regional · design partners (clinics + real estate) · voice + WhatsApp
 INDIA — Expand verticals + languages
   add edtech/restaurants/auto + Tamil/Telugu/etc · first paying customers · self-serve onboarding
 INDIA — Scale + optimize Indic AI
   tuned/self-hosted Indic models · multi-city · channel partners
 GLOBAL SPRINGBOARD
   reuse the platform for US/GCC · SOC 2 / ISO · India remains a strong base

Milestones

Foundation → Pilot: a tenant takes a real call in Hindi with a DLT-compliant number, AI disclosure, and a WhatsApp confirmation.
Pilot → Expand: strong call-containment rates and design partners willing to pay.
Expand → Scale: paying customers across two or more verticals and enough volume to justify optimizing the Indic model stack.
Scale → Global: repeatable onboarding plus published case studies.

9. India-specific risks & mitigations

Risk	Mitigation
Telephony / regulatory (PSTN, DLT)	Use a licensed Indian provider that handles DLT; verify media-streaming support early
Low willingness-to-pay	Affordable INR and per-call pricing; volume; WhatsApp bundle; efficient architecture
Language quality (accents, code-mix)	Indic specialists + per-language eval bench; human fallback
DPDP rules still finalizing	Conservative consent/notice now; track rules; counsel
Spam-label / reputation	Inbound and consented-transactional first; AUP; registered numbers
Fragmented integrations (local SaaS/PMS)	Integration tooling + WhatsApp + start with one or two popular vertical tools

10. India-first vs global plan — what changes (cheat sheet)

Topic	Global plan	India-first change (this doc)
Telephony	Global carrier	Licensed domestic India provider + DLT; global carrier for international only
AI models	Global ASR/LLM/TTS	+ Indic ASR/LLM/TTS for vernacular; tuned/self-hosted Indic models over time
Payments	Global gateways	India gateways + UPI links; UPI for in-call payments
Channel	SMS/email	WhatsApp-first + voice
Pricing	USD plans	INR plans / per-call, more accessible pricing
Compliance lead	GDPR / CCPA	DPDP Act + DLT + Telecom Act 2023
Entity	US parent	India Pvt Ltd (add a global parent at a global raise)
Languages	EN + a few	EN + Hindi + regional, code-mixed

Related: 07 — Vertical Opportunity Map · Back to 00 — Master Index

India Route Map