Why 200 Companies Are Using FireAI

Prakash's path to FireAI was not through foundation models or AI research. It came from years of running businesses and watching the same problem repeat itself across every industry he touched.

After graduating in electrical and electronics engineering in 2014, he worked at Zomato and BharatPe before joining his family's pharmaceutical business. In every role, the pattern was the same. Companies generated enormous amounts of data. Turning that data into decisions still required spreadsheets, delayed reports, and specialist analysts who took days to produce answers that were already outdated by the time they arrived.

The problem became most concrete when he joined the family business. Products would run out in one market while inventory sat idle in another, because sales, distributor, and warehouse data lived in entirely separate systems that never talked to each other.

That experience became FireAI's founding thesis.

Before writing a single line of production code, the founding team spent nearly five months speaking to businesses across sectors, validating pain points, and collecting enterprise datasets. Commercial operations began in November 2025 after multiple iterations with early customers. According to the company, nearly every major product feature came directly from customer feedback rather than an internal roadmap.

FireAI describes itself as a decision intelligence platform. The distinction matters.

Most business intelligence tools tell you what happened. A dashboard shows revenue is down. A report shows inventory is piling up. But explaining why it happened and suggesting what to do next requires a separate analyst, a separate process, and a separate set of tools.

FireAI is designed to close that gap. During a live demonstration, Prakash asked the platform why a company's sales were declining. Within seconds, the system traced the drop to its underlying business drivers, identified that nearly 40% of the decline came from a specific customer account, and generated a forecast of how the business could perform going forward.

Technically, the platform connects data across 700 or more sources, including ERPs, accounting software, CRMs, cloud databases, and marketing platforms. It creates a unified semantic layer without moving customer data outside their infrastructure, reading only database schemas and table structures rather than extracting complete datasets. SQL queries are generated directly against the customer's own databases, enabling real-time analysis while keeping data within the customer's environment.

The process works through a multi-agent architecture that sits before the language model. Specialised AI agents do the groundwork first: identifying relevant tables and fields, interpreting business context, mapping relationships between schemas, and defining the right formulas. Only after this preparation does the LLM generate SQL queries.

Under the hood, FireAI uses a fine-tuned version of Llama 3.3 for text-to-SQL generation, though the platform is model-agnostic and allows enterprises to integrate their own preferred models.

FireAI's multi-agent architecture reduces token consumption significantly because only structured context reaches the LLM. This allows the company to price through subscriptions rather than usage-based token billing, maintaining gross margins of around 80 to 85%.

The company also charges an upfront integration fee to connect enterprise data infrastructure, creating a two-part revenue model: integration fees to get started and recurring subscriptions for ongoing use.

The current client base includes 200 customers, with enterprises accounting for roughly 60% of revenue and MSMEs the remaining 40%. Beyond Bata and IRCTC, clients include Central Warehousing Corporation and a growing roster of mid-market businesses across manufacturing, logistics, and retail.

FireAI's early traction is real, but the path from Rs 9 Cr ARR to a scaled enterprise business is not straightforward.

Enterprise data remains fragmented. Many organisations still run on legacy ERPs, custom integrations, and siloed databases. Zero-friction onboarding is harder in practice than in the marketing promise, and it can take significant time before the platform's full value becomes visible to a customer.

Competition is intensifying. Global cloud providers are building more AI into their analytics stacks. Traditional business intelligence vendors are adding conversational features on top of existing dashboards. FireAI's task is to convince customers that it is not just another analytics tool, but a genuinely different way of working with data.

Enterprise sales cycles are long, management risk is real, and questions around control, accuracy, and data security are standard objections that every enterprise AI company has to work through carefully.

Despite these challenges, the direction of travel in enterprise technology is clear. The conversation is shifting from generating insights to enabling decisions. FireAI is betting that the real opportunity lies not in replacing business intelligence platforms but in building a reasoning layer on top of them.