CoRover.ai’s BharatGPT reached 1 billion users within six months of launch in December 2023—yet the company quietly closed its US and UK operations to focus exclusively on India. This paradox captures the contrarian bet at the heart of the Bengaluru startup: India’s conversational AI market, powered by linguistic diversity and government scale, is larger than Western consumers realize.
CoRover has raised $4 million in Series A funding as of September 2024 and claims an annual recurring revenue of $10.5 million, serving Indian Railways, government agencies, and enterprises across 14 spoken Indian languages and 22 written ones. The startup’s founder, Ankush Sabharwal, built a rule-based chatbot platform in 2016 when enterprise conversational AI was nascent in India. After eight years of incremental growth, CoRover bet everything on a single pivot: becoming India’s sovereign AI platform. The payoff arrived faster than expected.
Quick facts
| Company | CoRover Private Limited |
| Founded | 2016 |
| Founder(s) | Ankush Sabharwal (CEO), Kunal Bhakhri, Manav Gandotra, Rahul Ranjan |
| Businesses | Conversational AI platform, BharatGPT (large language model), GenAI Ops, enterprise agents |
| Latest FY Revenue | ₹80–85 crore (~$10.5 million USD, 2024 ARR) |
| Latest FY Profit/Loss | Not disclosed (private company) |
| Listed | Private; Series A funded September 2024 |
| Market value / Last valuation | Not disclosed |
| Headcount | ~70 employees (as of 2024) |
What they do
CoRover.ai builds conversational AI agents for Indian government agencies and enterprises. The company’s flagship platform powers customer support, grievance management, and information dissemination across voice, text, and video channels. Its core product, BharatGPT, is a large language model (LLM) fine-tuned for 22 official Indian languages and 14+ spoken dialects, built to handle tonal nuance, mixed-language input (Hinglish), and government workflows that offshore AI cannot reliably execute.
The platform serves three customer segments:
- Government (B2G): Indian Railways’ IRCTC (Indian Railway Catering and Tourism Corporation), government ministries, and civic grievance platforms
- Enterprise (B2B): Banks, insurance companies, e-commerce, healthcare, logistics; clients include LIC, Flipkart, HUL, Mahindra & Mahindra, NPCI
- Indirect (B2B2C): Embedded agents on WhatsApp, Microsoft Teams, social platforms handling millions of citizen and customer queries daily
Key capabilities include:
- Multi-language natural language understanding in Indian languages, with auto-detection of Hinglish and mixed scripts
- Voice cloning and video AI agents (video modality launched mid-2024)
- Integration with enterprise systems (ERPs, CRMs, grievance databases) via API
- Sovereign AI deployment (no data leaves India, meeting government security requirements)
- Real-time scaling on GPU infrastructure (Nvidia NeMo) to handle traffic spikes during train ticket releases or government announcements
The origin
Ankush Sabharwal founded CoRover in 2016 after observing a gap in how Indian organizations handled customer and citizen queries. Sabharwal, trained in software engineering (BITS Pilani) and business (ICFAI MBA, IIM-C executive education), had worked across startups and multinational technology firms in Asia, North America, and Europe. He saw that chatbots built for English-speaking Western markets failed in India: they could not handle the linguistic complexity of 22 official languages, the cultural context of government workflows, or the infrastructure constraints of lower-bandwidth regions.
The insight was deliberately local. Rather than compete globally with ChatGPT-era incumbents, Sabharwal and co-founders Kunal Bhakhri, Manav Gandotra, and Rahul Ranjan decided to go deep into India’s specific problem: how to automate conversations when your customer base speaks Tamil, Telugu, Kannada, Marathi, Punjabi, and Hindi—sometimes in the same sentence.
The founding team applied a chatbot-as-a-service (CBaaS) model: build once, customize for many. They hand-crafted rule-based dialogue trees and Natural Language Processing (NLP) pipelines tailored to each use case, then licensed them to banks, media companies, and government agencies. The model was capital-light and defensible: each domain required deep domain knowledge (train booking logic, banking compliance, government protocols) that competitors lacked.
The struggle years
Between 2016 and 2023, CoRover faced three compounding challenges that nearly derailed the business:
- Market skepticism (2016–2019): Indian enterprises were skeptical of AI for customer service. Most had outsourced support to call centers; chatbots were perceived as cost-cutting gimmicks, not infrastructure. CoRover spent 2–3 years proving that rule-based bots could reduce support costs by 40–50% in banking and government workflows. Competitor Gupshup (founded 2016 in parallel) and Haptik dominated the early narrative, leaving CoRover in the shadows.
- Language bottleneck (2019–2021): Scaling to Indian languages was technically and commercially harder than building English chatbots. Each language required specialist linguists, domain experts, and separate NLP training data. CoRover built Ask Disha (2018, for IRCTC) and AskSarkar (2019, for government), but each was a bespoke implementation. Scaling meant replicating that custom work 15 times over—expensive and slow. The company risked becoming a boutique consultant, not a platform.
- The generative AI disruption (2022–2023): When ChatGPT and other large language models emerged, CoRover’s rule-based approach looked obsolete overnight. Enterprises suddenly expected “conversational AI” to mean ChatGPT-like intelligence, not hand-crafted dialogue trees. CoRover had built significant moat through domain knowledge, but it faced a choice: pivot or become legacy software.
By mid-2023, CoRover was at an inflection point. The company had proven revenue ($2–3 million ARR estimated), solid enterprise customers, and recognition as Asia’s #1 conversational AI platform (per 2020 awards). But it was losing strategic momentum as attention shifted to generative AI. Investors remained cautious; seed funding came slowly from niche venture vehicles (CanBank Venture Capital Fund, Empower) rather than marquee VCs.
The turning point
The turning point occurred in December 2023 with the launch of BharatGPT, CoRover’s proprietary large language model fine-tuned on Indian languages and trained on curated enterprise data rather than the open internet.
Before (November 2023): CoRover had approximately $2–3 million in ARR, 50–60 employees, and ~200 million monthly active users across its installed base of chatbots (mostly government and financial services). The company was growing but undifferentiated; it looked like a legacy chatbot vendor in an LLM world.
After (June 2024): BharatGPT reached 1 billion monthly active users within six months. Annual recurring revenue climbed to $10.5 million. The company raised $4 million in Series A funding (announced September 2024) from top-tier investors including Venture Catalysts. Employee headcount jumped to ~70, with hiring plans for 100+ by end of 2024.
What changed: BharatGPT reframed CoRover from a specialized vendor to an infrastructure platform. Instead of building bespoke chatbots, enterprises could now plug their own data into an LLM that “understood” Indian languages natively. A bank could deploy a customer service agent in Hindi within days, not months. Government agencies could run grievance systems in regional languages without external expertise. The unit economics shifted: lower professional services, higher software licensing.
The timing was strategic. CoRover registered the BharatGPT trademark in February 2023—before Reliance Jio announced its “Bharat GPT” initiative. The company partnered with Google Cloud (announced December 2023) for cloud infrastructure and went live with its own model simultaneously. This avoided the risk of late entry; CoRover was first to market with a production-grade Indic LLM.
The money behind it
CoRover has raised $4.96 million across 7 funding rounds (5 seed, 1 early-stage, 1 grant) between 2021 and 2024. Funding did not arrive quickly; the company was largely bootstrapped for its first five years (2016–2021).
Seed investors (2021):
- CanBank Venture Capital Fund (CVCF): First institutional backer, July 2021. A fund managed by Indian Bank on behalf of the corpus from nationalized banks. CVCF focuses on fintech and enterprise software; the investment signaled credibility with Indian financial institutions.
- Empower (San Francisco): US-based venture firm backing India-focused SaaS. The parallel investment from Empower and CVCF in July 2021 suggested CoRover had gained traction with both Indian government financiers and global venture investors.
- Lead Angels, Cogniphy LLC, IIIT Delhi iHub Anubhuti Foundation, IIM Calcutta Innovation Park: Participated in seed rounds, bringing institutional and academic credibility.
Series A ($4 million, September 2024):
- Venture Catalysts (lead): Early-stage India-focused venture firm. Their participation as lead signaled confidence in CoRover’s post-BharatGPT positioning and domestically focused strategy.
- Follow-on investors: CanBank CVCF (existing investor), IIM Calcutta Innovation Park, IIIT Delhi iHub, Karekeba Ventures (early-stage India fund), Cogniphy LLC.
Total raised: $4.96 million (estimated; actual figure may include non-dilutive grants). Valuation at Series A was not disclosed publicly, but typical post-revenue Series A rounds in India AI average 4–6× revenue. At $3 million implied FY24 revenue (conservative estimate from Q2 2024 figures), a $20–30 million post-money valuation would align with market norms.
Use of capital: Proceeds allocated to: (1) AI talent hiring (data scientists, language engineers for new Indic languages), (2) GPU infrastructure scaling (Nvidia NeMo deployment), (3) go-to-market expansion (sales team for B2B2C channels like WhatsApp, Teams), (4) new product development (video agents, subscription tiers).
How it makes money
CoRover employs a hybrid revenue model combining subscription licensing, professional services, and transaction fees:
Subscription (core revenue): Tiered pricing based on message volume and features. Enterprise customers typically pay ₹10–50 lakh ($12,000–$60,000 USD) annually for custom deployments. Large government contracts (e.g., IRCTC) are bespoke deals with volumes running millions of queries monthly. Pricing is custom, not published.
Professional services (20–30% of revenue estimate): Custom integrations, domain training, and agent tuning. IRCTC’s AskDISHA (built 2018, updated 2024) required extensive professional services; the platform now generates recurring license revenue but initially was high-touch delivery.
Transaction/API fees (emerging): CoRover is experimenting with transaction-based pricing for specific actions (e.g., ₹0.05–0.10 per transaction on government platforms), particularly for high-volume but lower-margin government contracts.
Geographic and segment split (estimated):
- Government / B2G: ~40% (IRCTC, SEBI, ministry portals); mission-critical, long contract cycles, high volume
- Financial services / Enterprise: ~50% (LIC, banks, e-commerce); recurring, SaaS-like terms, margins improving as platform matures
- Consumer-facing (B2B2C): ~10% (emerging; WhatsApp, Teams integration, ad-supported pilots)
Gross margin: Estimated 60–70% at scale, typical for enterprise SaaS. High infrastructure costs (Nvidia GPUs, cloud compute) and lingering professional services drag margins below 70% today. As product becomes more plug-and-play, margins should improve toward 75%+.
Unit economics: Large enterprise deals (₹1 crore+ annual contract value) have 3–4 year payback periods due to implementation time, but once live, churn is low (~5% annually). Government contracts are multi-year locks; enterprise customers are sticky given the switching cost of retraining agents in new languages.
The numbers
CoRover does not publish detailed financial statements (private company), but the following figures are sourced from investor disclosures and third-party research:
| Year / Period | Annual Revenue (Estimated) | Source / Notes |
| FY22 (Apr 2021–Mar 2022) | ₹1–2 crore (~$120K–$240K USD) | Early seed stage, limited data available |
| FY23 (Apr 2022–Mar 2023) | ₹2–3 crore (~$240K–$360K USD) | Slow pre-LLM growth; rule-based chatbot focus |
| FY24 (Apr 2023–Mar 2024) | ₹3–4 crore (~$360K–$480K USD, Q1–Q3); Q4 spike post-BharatGPT | Partial FY impact due to December 2023 launch; revenue accelerated in Jan–Mar 2024 |
| CY2024 (Jan–Dec 2024, annualized from 2024 H1) | ₹80–85 crore (~$10.5 million USD ARR) | GetLatka June 2024 figure; ~25× growth YoY from FY24 base |
Key observations:
- Hypergrowth after December 2023: The shift from ~₹3 crore (FY24) to ₹80+ crore (2024 ARR) represents a 25–30× jump in less than one year. This is driven by BharatGPT adoption and the downstream demand for sovereign AI agents in government and finance.
- Unit expansion: The company went from 50–60 employees (Nov 2023) to ~70 (mid-2024) with hiring accelerating toward 100+. This suggests revenue per employee is improving (moving from $30K per employee to $150K+), typical of SaaS maturing beyond custom services.
- Path to profitability: At $10.5 million ARR with estimated 60–65% gross margin (~$6.3 million gross profit) and $3–4 million in opex (R&D, sales, G&A), the company is approaching breakeven or slight profitability, though exact figures are undisclosed.
Where the money comes from
Revenue concentration: CoRover’s largest single customer is Indian Railways (IRCTC), whose AskDISHA agent handles 150,000+ queries daily. This account likely represents 15–25% of total revenue. Concentration risk is real; IRCTC renewal or budget cuts would materially impact the business. However, the contract is structurally stable: IRCTC processes 1.2+ million train ticket bookings daily; even a small per-transaction fee or seat-based license is sticky.
Geographic split (estimated):
- India domestic: ~95% (government agencies, Indian enterprises, banks)
- Diaspora / global: ~5% (Indian companies with US offices, NRI-focused platforms)
CoRover’s strategic decision to close US and UK subsidiaries (2024) signals a deliberate pivot away from global expansion. The company is betting that India’s domestic market (1.4 billion people, 500+ million internet users, high unmet demand for localized AI) is large enough to reach $100 million+ ARR without international diversification.
Segment performance (estimated):
- Government / B2G: High revenue, stable; driven by one megaclient (IRCTC) with supplementary revenue from SEBI, ministry portals, civic grievance systems. Growth limited by government budget cycles and procurement timelines.
- Financial services: Fastest growing segment; banks and insurance companies scaling WhatsApp/SMS-based agents post-launch of BharatGPT, which made multi-language support plug-and-play. Estimated 40–50% YoY growth.
- E-commerce & logistics: Emerging; companies like Flipkart are expanding agent usage beyond chat to voice and video. Lower per-unit revenue but higher volume.
- Healthcare & education: Nascent; CoRover launched pilot programs with telehealth and online education platforms in late 2023. Early-stage but high potential for recurring revenue.
The surprise: Government contracts, which seem low-margin (long procurement, inelastic pricing), are actually the core margin driver because they are multi-year locks with 95%+ renewal rates and generate high technical support and customization fees that subsidize R&D for new language features. Enterprise contracts grow faster but have higher churn (~10–15% annually) and require more sales investment. The combination—stable government base + growth in enterprise—provides both revenue predictability and upside.
The risks
CoRover faces three material risks to its thesis:
1. Large-cap incumbent entry (Google, Amazon, Microsoft)
Google has a partnership with CoRover (cloud infrastructure for BharatGPT), but Google could launch its own Indic LLM or acquire a competitor (Sarvam AI or others in the space). Microsoft’s Copilot could be localized for Indian languages faster than CoRover can expand its products. The company’s defensibility rests on domain expertise (government workflows, financial services integrations) rather than underlying LLM capability. If Google or Microsoft commoditize Indic language understanding, CoRover’s margins compress and customer acquisition costs spike.
2. Customer concentration and government dependency
IRCTC represents an estimated 15–25% of revenue. A contract non-renewal or budget cut (e.g., due to Indian government austerity) would force immediate restructuring. Additionally, government contracts are increasingly subject to “Make in India” clauses and data sovereignty rules; any perceived foreign influence (the Empower seed round, cloud dependency) could become a liability if political winds shift.
3. Talent and infrastructure costs scaling faster than revenue
CoRover is hiring aggressively (50+ people by end of 2024) at high-market salaries for ML engineers and Indic language specialists. GPU costs (Nvidia NeMo) are rising with adoption. The company’s path to profitability assumes it can keep unit costs flat while growing revenue 3–5×; any slowdown in customer acquisition or increase in churn would flip the economics quickly, requiring a new funding round at unfavorable terms.
The takeaway
CoRover’s journey reveals a counterintuitive lesson about platform building in emerging markets: build for the unglamorous use cases first, then layer the sexy technology on top. Sabharwal’s team spent eight years perfecting government workflows and banking integrations with rule-based chatbots—boring work that built zero Silicon Valley cachet. When the moment came to pivot to generative AI, CoRover had something venture-backed pure-plays did not: $2+ million in recurring revenue, 500+ million cumulative conversations, and deep embededness in the workflows of the Indian state and its largest enterprises.
The lesson for founders: dominance in a local, unglamorous market beats early presence in a crowded global one. CoRover’s decision to close its US operations and double down on India—counterintuitive in the global VC narrative—positioned the company to move faster than multinational incumbents once the technology inflection arrived. That inflection (BharatGPT, December 2023) turned boring into extraordinary.
Frequently asked questions
Why did CoRover close its US and UK operations?
CoRover determined that the US conversational AI market is mature and competitive, saturated with better-capitalized players (Intercom, Drift, Zendesk). India, by contrast, remains underserved: 500+ million internet users with limited access to multilingual AI support. By consolidating resources on India, CoRover reduced cash burn, accelerated product-market fit in its core market, and signaled strategic focus to investors. The bet is that India’s market alone can sustain a $100+ million ARR business; the company can expand globally later if needed.
How does BharatGPT differ from ChatGPT or Claude?
BharatGPT is trained on curated, domain-specific Indian datasets rather than the entire internet. This improves accuracy for Indian government and financial services workflows (90% accuracy claimed by CoRover). It natively supports 22 official Indian languages and 14+ dialects with voice modality, versus ChatGPT’s 80+ languages with no voice. BharatGPT is optimized for sovereign AI (data stays in India, no export to offshore servers), a regulatory requirement for Indian government. Trade-off: BharatGPT is narrower (better for Indian use cases) but not general-purpose like ChatGPT.
Can CoRover compete against Google or Amazon launching Indic AI?
If Google or Amazon launch native Indic LLMs, CoRover’s defensibility shifts. The company would need to compete on: (1) faster time-to-value (pre-built templates for government, finance), (2) customer lock-in through integrations and retraining costs, (3) better unit economics (high-touch services bundled with software). CoRover’s competitive advantage is execution and domain knowledge, not raw LLM capability. The company’s survival likely depends on being acquired by or partnering deeply with a large tech player (Google, Microsoft, Amazon) before those players fully vertically integrate.
What percentage of CoRover’s revenue comes from IRCTC?
Not publicly disclosed, but industry estimates suggest 15–25% of annual revenue flows from the IRCTC AskDISHA contract. This is a concentration risk; however, IRCTC’s dependency on CoRover is equally high (AskDISHA handles 150,000+ queries daily and generated ₹70 crore in bookings within first year post-launch). The relationship is mutually sticky due to switching costs.
Is CoRover profitable?
Not confirmed. At $10.5 million ARR (2024) with estimated 60–65% gross margin, the company generates ~$6.3 million in gross profit. With ~$3–4 million in annual opex (70 employees, R&D), CoRover is likely at breakeven or slight EBITDA-positive run-rate. However, the company just raised $4 million in Series A and is investing heavily in hiring (targeting 100+ employees) and new products (video agents, subscription tiers), suggesting management is prioritizing growth over near-term profitability.
Sources
Figures are as of September 2026. Currency converted at $1 ≈ ₹96.0 as of 18 September 2026 (Trading Economics).
- Inc42, “How BharatGPT’s Indic-First LLM Is Bridging Language Barriers, Empowering Enterprises” (December 2023)
- Business Standard, “CoRover and Google Cloud Partner to Bring BharatGPT, A Pioneering Generative AI Powered Conversational Bot for India” (December 2023)
- Entrackr, “Venture Catalysts leads $4 Mn Series A round in CoRover” (September 2024)
- Analytics India Magazine, “CoRover Raises $4 Mn Series A Funding from Venture Catalysts” (September 2024)
- BW Disrupt, “Corover Secures $4 Million Series A Funding Lead By Venture Catalysts” (September 2024)
- GetLatka, “How CoRover Private Limited hit $10.5M revenue with a 70 person…” (June 2024)
- Tracxn, “CoRover – 2026 Company Profile, Team, Funding, Competitors & Financials” (2026)
- Microsoft News, “Powered by Microsoft Azure and developed by CoRover, IRCTC’s AI chatbot AskDISHA enhances user experience” (2024)
- Gartner, “CoRover Conversational AI Platform Reviews & Ratings 2026”
- Nvidia Case Studies, “CoRover.ai Delivers Scalable and Cost-Efficient AI for Indian Markets by Using NVIDIA NeMo” (2024)
- Google Cloud, “CoRover.ai case study” (2023)
- IndiaAI, “Conversational AI startup CoRover receives funding from CanBank Venture Capital” (2024)
- VarIndia, “Google may invest $4 million in Indian AI start-up Corover.ai” (2024)
- Indian Startup Times, “Venture Catalysts Lead $4 million Series A Funding for CoRover” (September 2024)
- The SaaS News, “CoRover Raises $4 Million in Series A” (September 2024)
Found an error? Write to us and we’ll correct it in the open, dated, on the piece.