As India’s BFSI sector accelerates its digital transformation, artificial intelligence is emerging as the backbone of modern lending infrastructure. From real-time underwriting to fraud detection, AI is not just enhancing legacy systems, it’s redefining them. One company at the forefront of this shift is Recur Club, a revenue-based financing platform that operates as a debt marketplace for recurring-revenue businesses. Backed by a tech-first DNA, Recur Club is reimagining how capital flows in the startup and SME ecosystem. In a conversation with ObserveNow Media, the company’s Chief Technology Officer, Anirudh Bhardwaj, shared insights into how AI is helping the platform scale securely and intelligently.

While traditional lending still hinges on human judgment and fragmented datasets, players like Recur Club are embracing AI-native workflows. Bhardwaj reveals that their approach begins with real-time parsing of financial statements and compliance data using large language models (LLMs). 

1. As a debt marketplace, what specific AI use cases have you deployed in your risk assessment or underwriting process?

At Recur Club, we’ve reimagined risk assessment using AI-first principles. We’ve deployed LLM-powered document intelligence to parse and analyze financial statements, bank data, and compliance documents in real-time. Our underwriting agents can identify anomalies, compute risk scores, and recommend credit decisions faster than any manual process. Additionally, autonomous loan agents nudge borrowers to complete data gaps, enhancing input quality for downstream models.

2. Beyond financial metrics, what key performance indicators (KPIs) do you prioritize to measure success and ensure long-term sustainability?

We track deal velocity (lead-to-deal and deal-to-funding turnaround), data completeness score, and borrower experience index as leading indicators. These reflect operational health beyond just disbursed value. Additionally, AI adoption rate per workflow, and percentage of autonomous completions help us ensure the platform is scaling sustainably, not just growing.

3. Can you share your approach to combating financial fraud using AI or analytics?

Fraud detection in lending demands context-aware intelligence. We’ve trained our AI models on borrower behavior patterns across hundreds of data points—bank transactions, MCA filings, GST data—to catch inconsistencies and potential manipulation. We also use deep fuzzy matching to catch shell entities and OCR+AI combinations to validate document authenticity. This multilayered approach flags potential fraud before funds leave the system.

4. What are the biggest challenges you foresee in the BFSI sector as of now and how many have we overcome in the new?

The biggest challenge is data fragmentation—most BFSI players still operate in silos. AI can’t thrive on patchy or delayed data. We’ve partly overcome this through automated data orchestration pipelines, ensuring near real-time ingestion of compliance and financial data. However, the broader challenge remains: building interoperable, AI-ready infra across the lending ecosystem.

5. Have you integrated blockchain into any part of your operations? If yes, where and how?

While blockchain isn’t part of our core underwriting yet, we’re actively exploring it for tamper-proof audit trails of document verification and for tokenizing receivables to enable smart contract-based debt instruments. Our current pilots focus on lending syndication records and ensuring traceability across multiple stakeholder actions.

6. Have you adopted any AI-driven KYC or fraud detection systems? What has been the impact?

Yes. Our AI agents automate entity extraction and verification from KYC documents, match PAN/Aadhaar/MCA data, and flag edge cases in real-time. Impact? KYC TAT reduced by over 60%, manual dependencies are down, and the system self-learns from previous exceptions. This has improved trust with lenders and borrowers alike.