AI-BPO: The SubVerse AI Interview
Can AI + Voice Solve Real Life Problems like a human customer care agent? IIT-Bombay grads Rishi Kumar, Tanmay Lad, and Nilkanth Parmar are building SubVerse--an AI BPO focused on BFSI market.
It was supposed to be a controlled demonstration.
Tanmay Lad, Rishi Kumar Murugan, and Nilkanth Parmar, co-founders of SubVerse AI, had traveled to the Hitachi Payments office in Chennai to showcase their AI voice agent technology to the company's CTO.
After a lengthy meeting discussing potential use cases, the plan was to set up a controlled demonstration: a conference call where their AI voice agent and one of Hitachi's human customer service representatives would both be on the line with an actual customer.
What happened next was entirely unplanned—and completely changed the trajectory of their relationship with Hitachi Payments.
“While setting up that call, the chief engineer was busy and he was trying to fix it, but he just messed it up,” Rishi recalls with candid humor.
“He was supposed to arrange a conference call with the voice agent and one of their customer service agents. They were both supposed to be in the same call with the actual customer.”
In the scramble to set up the demonstration, something went wrong. The human agent never made it onto the call. It was just the customer and SubVerse’s AI voice agent, ‘Pooja’. Hitachi Payments was handling the call on behalf of its client SBI Payments.
Here is the raw call recording without any edits. Hit the play button on the right:
Rishi explains:
“The customer started explaining the problem they were facing. The AI completely took over the call and resolved the issue on its own.”
The AI voice agent listened patiently as the merchant, Kingsley, explained the socket error he was experiencing with his POS terminal. Pooja then guided him step-by-step through the troubleshooting process.
The AI agent even handled an additional request for paper rolls, creating a delivery order and providing a tracking number:
“Your complaint ID is F for France, V for Varanasi 2383462. I will also send you a WhatsApp message for tracking.”
The customer never realized Pooja was an AI agent. The AI's performance left the Hitachi Payments team stunned.
“By the time the chief engineer's panic subsided, the whole call had already finished. He was searching through the database to see what had happened during the call,” Nilkanth recounts.
What truly impressed the Hitachi CTO wasn't just the seamless problem resolution, but the AI's linguistic adaptability.
“It even switched languages between English and Tamil. The customer was speaking in two different languages,” Tanmay explains.
“The CTO was actually flummoxed by what had happened. He was so happy that he took us out for dinner that evening. And that's how we secured our partnership with Hitachi Payments.”
This unplanned, real-world test had demonstrated something crucial: SubVerse AI's voice agents weren't just capable of handling scripted demos—they could manage authentic customer interactions under pressure, delivering faster resolution times than human agents while maintaining a natural, respectful conversation flow.
Background
Tanmay Lad grew up in Jalgaon in a small business family where he learned the basics of business by helping his family run a stationary shop.
Meanwhile, Rishi Kumar was raised in Mumbai’s Income Tax quarters—his father has been a career Income Tax Officer for 35+ years, due to retire in 2025.
Tanmay and Rishi met each other on their first day on the IIT-Bombay campus—they were roommates.
After graduation, they initially took different career paths. Tanmay joined Cairns, an oil and gas company under Vedanta, where he excelled.
In one notable achievement, he found an innovative solution when four oil wells converted to polymer wells were sitting idle due to commissioning delays. By modifying the process using in-house inventory, Tanmay's initiative resulted in the production of approximately 100,000 barrels (valued at $6.8 million) with minimal extra cost—earning recognition from top management.
Later, he pursued an MBA at IIM Ahmedabad, excelling academically and winning prestigious scholarships like:
OP Jindal Scholarship (1 out of 10 selected from 389 students) and the Aditya Birla Group Scholarship (1 out of 20 from 389).
He then joined BCG, where he led a project to revamp field sales for a major public sector bank, improving salesforce effectiveness by 30% in just three months.
Rishi was a founding member of Lovetreats, an online sexual wellness products retailer where he creatively addressed cultural taboos.
He recounts:
“While building Lovetreats we figured India has a lot of taboos surrounding sexual wellness products delivered online at home (joint family households). So, we registered the company as LT Digital Private Limited and had that as the shipping address on the packaging.”
This approach led to the highest percentage of female users (approximately 65%) among competitors. Impressively, he helped the company grow completely bootstrapped to $1 million in revenue within 18 months of launching.
He then also spent time working with Zoomcar to build PEDL, a last-mile rental bicycle program launched across multiple cities at an affordable rate of 20 rupees per hour.

Nilkanth Parmar, the third co-founder who would join them later, started his freelancing journey after graduating from NIT Surat. He landed clients from the US, Spain, Australia, etc., slowly scaling his team in the process.
He had worked on numerous tech projects, including the open-source project AsyncAPI/Modelina, which creates data models in any language based on an Event Driven Schema; MployTrakr, an HRMS system for MSMEs; Ellu, a Spanish-based E2E encrypted chat platform; and several other significant projects across various tech stacks.
Opportunity
By August 2022, Tanmay, Nilkanth, and Rishi had come together to start something meaningful together.
Tanmay had realized that consulting wasn't creating real-world value:
“The kind of value that we were creating is just in PPT and people's minds. That value is not actually created unless and until people at the client side actually work upon it and execute on it.”
They spent months researching various industries—from electric vehicles to D2C brands to import-export—before eventually zeroing in on the potential of AI, particularly large language models (LLMs).
By April 2023, they recognized that the core technologies they were working with—text-to-speech, speech-to-text, and LLMs—could be applied to create voice agents for customer service, with particular potential in the banking, financial services, and insurance (BFSI) sector in India.
With Tanmay and Nilkanth leading AI product development and backend architecture, and Rishi focusing on sales, marketing, and business development, they had become a team with complementary skills.
The Struggle to Land the First Big Customer—Acko
Securing the first major client wasn't easy.
Even when potential clients showed interest, the path to implementation was filled with obstacles, particularly around integrations with existing systems.
“We got in touch with one of the ex-CXOs through LinkedIn and explained what we were trying to build,” Rishi explains. “He then set up a call with one of his former teammates who was managing customer service.”
After a successful demo, they moved to a ‘paid dogfooding’ process where they demonstrated how their voice agent would handle actual customer calls.
Here is a raw unedited call recording of the SubVerse AI agent talking to a customer for policy renewal. Hit the play button on the right:
“The person was surprised by what was happening—wondering who this agent was that was calling about policy renewal,” Rishi recalls.
The guy was so convinced he was speaking to a human that "he started asking questions like ‘which city are you from?’ and ‘how many years have you worked for Acko?’”
The transcript of this call shows just how convincingly human the AI appeared:
When the customer asked, "Where are you from, Pooja?" the AI gracefully deflected: “I am glad you are curious, Mayankji. But my main goal is to help you with your insurance needs.”
Even when directly challenged with “Are you a human or are you a bot?” the AI maintained its composure: “I understand the confusion, Mayank-ji. I am a virtual assistant here to help you with your insurance queries.”
What made the AI so convincing wasn't just the natural-sounding voice but its comprehensive capabilities:
“That's the level of quality we've achieved with our voice agent—it's human-like, always patient and responsive to how customers speak. It handles interruptions naturally and has all the relevant details about both the car and the customer.”
Acko is now a paying customer of SubVerse AI.
These early successes with Acko and Hitachi Payments (SBI Payments) validated their approach. But as they worked deeper with Acko and Hitachi Payments, they realized that voice agents alone weren't enough—clients wanted comprehensive solutions that could handle various tasks and integrate with their existing systems.
The Evolution of the Product
What began as an AI voice agent evolved into a complete workflow automation platform.
“Just having a voice bot isn't enough—the voice agent is only the front-facing component of what clients really need,” Tanmay explains.
“Behind that, you need specialized agents for different tasks, all working together seamlessly.”
They expanded beyond voice to include Email and WhatsApp agents that could interact with a company's knowledge base to answer queries, create tickets, assign them to relevant departments, and check compliance against company policies.
The core innovation was making these complex workflows simple to design.
Rishi notes:
“Traditional SaaS platforms are too static for this purpose. With our system, you simply provide broad instructions like 'do this, do that,' and the platform handles everything else.”
The system intelligently determines which specialized agent to deploy at each moment, based on the actual content of the conversation.
The transformative impact of their technology was evident in the metrics. Traditional IVR systems had dismal engagement rates.
Tanmay explains:
“Less than 10% of people listen to the entire message; most hang up within 10 seconds. In contrast, with SubVerse’s voice agents, more than 60-70% of people engage in actual conversation, which is comparable to human-agent interactions.”
Nilkanth adds:
“Our AI agents speak multiple languages in natural, human-like voices. They handle complex queries and integrate seamlessly with existing systems. They are always available, delivering consistent, and quality service.”
BFSI Industry-Specific Solution
As their product gained traction, SubVerse AI secured funding through an incubation program at SINE, IIT Bombay.
Currently, both the BFSI enterprise clients of SubVerse AI are processing around a few hundred calls per day and around a thousand workflow operation tasks.

While Acko is already converting into a long-term contract, Hitachi Payments which started a few months after Acko is moving quickly toward a long-term conversion.
Through their journey, Tanmay, Rishi, and Nilkanth have demonstrated that combining complementary skills—Tanmay's business acumen and technical prowess, Rishi's entrepreneurial drive and market understanding, and Nilkanth's technical development expertise—can create solutions that fundamentally transform how businesses interact with their customers.
The accidental demonstration at Hitachi Payments wasn't just a lucky break—it was the culmination of a deep understanding of both technology and market needs. As Tanmay concludes in their pitch:
“Customer service is unncessarily complicated—we want to work with BFSI clients and fix it in the long term for the industry as a whole.”
SubVerse AI is beginning its pre-seed capital raise, where they are aiming to raise $500K from industry-specific angels and funds.
If you want to work with the SubVerse AI team or want to know more details about investing in the company, please email us at banjan@tal64.com.