Problem
Indian SMEs — real estate brokers, insurance agents, coaching institutes, and D2C brands — generate thousands of leads via Meta ads and IndiaMART but lack the manpower to follow up within the critical first hour. Manual dialler teams cost ₹25,000–40,000/month per seat, and most leads go cold before a human ever calls. Conversion on unworked leads averages under 2% in India.
Solution
A web-based platform where an SME uploads a lead CSV (or connects a Facebook Lead Ads webhook), selects a Hinglish or regional-language voice persona, and the AI agent begins calling within minutes. The agent introduces the business, qualifies intent (budget, timeline, location), and either books a callback slot in the founder's calendar or pushes a WhatsApp transcript of the conversation. No hardware, no call centre setup — live in under 30 minutes.
Why Now
Inc42's May 2026 analysis identifies AI vertical SaaS as India's single fastest-growing startup segment, with enterprise automation drawing the sharpest investor interest. Simultaneously, Indian-language voice APIs (Sarvam AI for TTS/ASR) and PSTN connectors (Exotel) have reached production maturity at sub-₹0.50/minute pricing — making per-call economics viable for SME budgets for the first time. US counterparts (Bland AI, Vapi YC W23, Retell AI) achieved product-market-fit in 2024–2025, validating the playbook globally before India adoption.
Target User
First 1,000 customers: real estate channel partners in Mumbai, Pune, Bengaluru, and Hyderabad running 200–1,000 leads/month from Meta or 99acres. Monthly revenue ₹3–5 lakh per broker team, so ₹10,000/month SaaS is a trivial line item. Purchase trigger: a weekend campaign that generated 400 leads, of which 380 went uncontacted by Tuesday.
Business Model
Monthly SaaS: Starter ₹5,000 (1,000 minutes), Growth ₹12,000 (3,000 minutes), Pro ₹25,000 (10,000 minutes + CRM sync). Overage at ₹1.20/minute. COGS: Exotel ₹0.40/min + Sarvam AI TTS ₹0.10/min + infra ₹0.05/min = ₹0.55/min total. Gross margin 55–65%. Break-even at 80 paying customers (≈₹6.4L MRR).
Competitive Landscape
- Direct (India): Yellow.ai and Kore.ai offer voice bots but target enterprise (₹5L+ ACV, 3-month onboarding). Ozonetel provides IVR infrastructure but not conversational AI agents. No SME-focused outbound AI caller product exists in India.
- Direct (global reference): Bland AI (US, $72M raised 2025), Vapi (YC W23) — both US-only, no Indian language support, pricing unworkable for Indian SMEs.
- Why we win: Hinglish-native voice personas, Exotel/Indian PSTN integration, sub-₹10,000/month pricing, and 30-minute self-serve onboarding vs. months of enterprise implementation.
6-Month Plan
- Month 1 (₹1.5L): Core integration — Exotel outbound call flow, Sarvam AI TTS, lead CSV ingestion, single real estate qualification script. Deploy to 5 beta partners.
- Month 2 (₹1L): Live call transcription, WhatsApp summary via Twilio, Calendly webhook for booking. Iterate on voice quality from beta feedback.
- Month 3 (₹1.5L): Self-serve onboarding, Razorpay subscription billing, persona script builder. Launch to 50 paying customers.
- Month 4 (₹1.5L): Facebook Lead Ads webhook, IndiaMART lead connector, Zoho CRM sync. Expand to insurance and coaching verticals.
- Month 5–6 (₹2.5L): Multilingual personas (Tamil, Telugu, Bengali). Conversion analytics dashboard. Target 150 paying customers.
Total capex: ₹8L. Remaining ₹8L is 4-month burn runway.
Risks
- TRAI DNC regulation: TRAI's Do-Not-Call registry and commercial communication rules could restrict AI outbound calls; mitigate by building mandatory DNC scrubbing into the product and positioning strictly as "opted-in lead follow-up."
- Voice quality churn: If the AI sounds robotic, SMEs churn in week one; mitigate by using Sarvam AI's natural Indian-English TTS and running script A/B tests with beta cohort before general launch.
- Yellow.ai moving down-market: Large players could launch SME tiers; mitigate by building deep vertical integrations (99acres, Housing.com, insurance aggregator APIs) that take 6+ months for incumbents to replicate.
Score Breakdown
Market (17/20): 6+ crore Indian MSMEs, with 500K+ active outbound sellers (real estate, insurance, coaching, D2C) representing a ₹1,500Cr+ annual TAM at ₹5,000–15,000/month ARPU — clearly above the ₹1,000Cr threshold for full marks.
Capital (13/15): MVP ships for ₹5–7L using fully API-first architecture with no proprietary ML; 12-month runway fits within ₹8L total capex — well inside the ₹15L comfort zone.
Team (8/10): One full-stack developer and a founder-salesperson can ship v1 in 6–7 weeks; a third contractor for prompt engineering is optional. All complexity is integration work, not original research.
Trend (13/15): Inc42 May 2026 names AI vertical SaaS as India's hottest startup segment; US voice AI agents (Bland AI, Vapi) validated demand globally in 2024–2025. India-specific pricing and language gap remains wide open.
Moat (10/15): Near-term moat is Hinglish-native personas and Indian telephony depth; medium-term moat is a call-data flywheel improving conversion-rate prediction per vertical. Not deeply defensible until 12+ months of data accumulates.
Economics (12/15): 55–65% gross margins with viral word-of-mouth distribution through broker WhatsApp groups. CAC estimated ₹3,000–5,000 via founder-led sales; LTV at 18-month average tenure is ₹90,000–2,16,000 — healthy 20–40× LTV:CAC ratio.
Speed (8/10): First paying user achievable in week 7; self-serve public launch by week 10–12. Slightly above the ≤6-week ideal but well inside the 12-week benchmark.