Startup Mantra: AI is powering document automation

Hindustan Times

02-08-2025

Rahul Bhanose, founder and chief executive officer (CEO) of bizAmica, had worked primarily with product companies, most recently with a US-based company, One Network Enterprises, where he was hired to set up a Global Capability Centre (GCC) in Pune. Says he, 'Here I learnt all one needs to know to become an entrepreneur. We started with one person, and I built up a team of 50 in 15 months. I looked at every aspect of the business and decided that I wanted to set up my own company.' bizAmica co-founders (from left) Rahul Bhanose, Harshada Bhanose and Manoj Godse. (HT PHOTO)
Initial steps
In 2011, he established a company in the education sector as a video-based learning platform. But after three and a half years, he shut that down. Later, in 2018, he saw that document processing was a big problem that needed a solution. Industry data showed that financial institutions process millions of documents annually, with volumes continuing to surge despite digitisation efforts. The persistence of paper-based process flows stems from several factors: entrenched legacy processes, customer preferences for physical documentation, regulatory requirements mandating signed or notarised forms, and the ongoing need to digitise historical records.
Says Rahul, 'Despite substantial investments in digital platforms and online customer portals, Banking, financial services, and insurance (BFSI) institutions remain heavily dependent on document-based processes. From customer onboarding and loan origination to claims processing and regulatory compliance, critical business functions continue to revolve around physical and digital documents spanning multiple formats—scanned images, PDFs, emails, and handwritten forms—often scattered across various business processes.
The document dependency is to address the financial and regulatory risks that have created a significant operational burden.' That needed a solution.
Idea to product
Rahul set up a team of four people to find a solution to this. Before venturing into document processing, Rahul had created an AI-based multilingual voice bot that had a use case in call centres. Call centres at banks could be handled by their AI voice bots, and only if this tech tool could not handle a conversation, it went to a human. But it was perceived as a saturated market.
So, Rahul and team built izDOX AI Platform. 'The most formidable obstacle for a bank is the handling of unstructured and semi-structured data that is embedded within documents. While structured data from transaction logs and Customer Relationship Management (CRM) systems can be readily processed, unstructured content—including customer data, digital forms, and handwritten forms—demands sophisticated processing capabilities.
'Traditional Optical Character Recognition (OCR) systems, the industry's go-to solution for document processing, frequently fail to deliver the required results for mission-critical applications. These systems struggle particularly with low-quality scans, handwritten text, and multilingual content, forcing institutions to fall back on manual 'stare and compare' processes where staff must visually review and validate documents—a methodology that is both time-intensive, stressful, and error-prone.
With an investment of ₹ 1 lakh, Rahul started bizAmica. Says he, 'While developing an AI tool, it is most important that you have a person with domain knowledge. With our existing domain knowledge from the BFSI industry and our technical expertise, we built the izDOX AI platform.
'The BFSI industry carries a 6% chance of fraud. We trained our tool to identify those risks and inconsistencies whilst processing BFSI documents.'
To the market
During Covid, Rahul got introduced to ICICI Lombard. 'Large institutions are difficult to penetrate, especially when you are a startup. We met ICICI Lombard as our first customer through a referral. They were impressed with our offering and signed us during Covid.'
Post that, Rahul knew that accessing large enterprises wasn't easy. So, he drew up an ideal customer profile that was largely the chief technical officer (CTO), chief information officer (CIO) of the BFSI industry. To reach out to them, he participated in events and set up booths where people could see what they were offering. 'We had a fairly good response. 75% of the people who showed interest asked for a commercial proposal.' However, conversion was another story and meeting people at events fairly limited their scope.
Realising that meeting the Ideal Customer Profile (ICP) in large enterprises was a long process, Rahul devised another strategy. He decided to tie up with organisations that are offering a solution or services to banks. He built partnerships with them. 'We have four types of partners: system integrators, robotic process automation (RPA) providers, low-code & no-code providers, BPO services for banks and lastly sales channels.' Of these, the RPAs and BPO back-office channels worked most efficiently. As of now, he has 15 customers in India.
Funding
Rahul was very keen on being cash positive. 'I did not want to borrow any funds when we were developing our product, hence we took product development assignments from US-based companies. That helped us fund our product development.' This served us well. The money that we earned through product building for them funded our izDOX AI platform journey.
But now that the company has proven its product market fit, he plans to reach out to markets in the Middle East, Africa and Southeast Asia. 'For this, we will need more funds. So far, we have managed on our own. Our revenues this year have been ₹4 crore, and in this financial year we aim to grow 2.5 to 3x. But expanding geographies will need more money. We plan to approach venture capital (VCs) for that.'