The key lesson for Piramal Finance is simple: being AI-native is not about an AI model or one team. It’s an ecosystem where every stakeholder, from data scientists and AI Manufacturers to technology and business users, has a critical role to plan, says Markandey Upadhyay, CDAO, Piramal Finance.
Markandey Upadhyay, Chief Data and Analytics officer, Piramal Finance.
Your organisation uses a wide range of AI and agentic solutions, across productivity, risk management, customer experience, and revenue. How do you prioritize which use cases to scale first, and what lessons have you learned in moving from pilots to enterprise-wide deployment?
Our approach to AI is very clear, we build it from the ground up, never force-fitting AI into a business solution. We call this our AI-native ecosystem, which consists of three pillars: business users who are ready to consume AI, our ability to manufacture high-quality models, and the technology that makes these models usable in the critical path of decision-making.
Because business users are involved from the start, every solution goes through multiple rounds of iteration—training supervised models, fine-tuning the model as per user feedback, refining the interface, and making adoption natural. That’s how we move from 10% usage to near-universal adoption.
The key lesson for us is simple: being AI-native is not about an AI model or one team. It’s an ecosystem where every stakeholder—from data scientists and AI Manufacturers to technology and business users—has a critical role to plan.
AI coding agents, hiring assistants, and login desk elimination point to a transformation in how employees work. How are you preparing Piramal’s workforce and culture for this shift toward AI-augmented role?
These innovations are not just about technology, they represent a cultural shift. While the technology was available to use, what really enabled us to start getting many of the projects off the ground, is the AI-native culture we have inculcated here at Piramal.
A lot of the demand for our projects is directly coming from the end users. Once a team sees the benefits, they quickly ask for upgrades and spread the word across departments. In fact, we are faced with the positive challenge of how to allocate resources to various AI initiatives.
Additionally, we have focused on the training of our staff. We have created infrastructure, where staff, with no prior coding experience, are capable of making their own small tech solution. As long as these systems enable people to do more, do better and make their job easier, I feel there is a natural gravity towards AI used systems.
What was the core strategic vision behind deploying AI and agentic solutions across Piramal Finance?
At Piramal Finance, becoming an AI-native financial services company is at the heart of our strategic vision. We see AI not as an add-on, but as the engine driving growth, sharper underwriting, higher productivity, and a superior customer experience.
To deliver this, we are investing deeply in in-house talent, best-in-class cloud infrastructure, and advanced technologies such as supervised deep learning models, generative models, MCP and agent-to-agent frameworks. These allow us to build multi-agent solutions that solve real business problems at scale.
Our vision is clear: AI must be our competitive advantage, enabling seamless growth while keeping risk under control and improving profitability.
With tools like Arya, the AI assistant for frontline employees, and AI-driven customer interactions, how do you ensure the right balance between automation and the human touch, especially in a customer-centric business-like financial services?
As a financial services company, human interactions are extremely important to us. We call ourselves high-touch and high-tech. Our approach is human centric.
We use AI to empower our employees to make better and faster decisions. Lending, especially for things like housing, will always require a human touch. It is one of the largest financial decisions that most Indians will make. Our AI solutions are built with an objective of trying to enable more intelligent human interactions, while AI synthesises away tedious tasks.
For example, searching on which page a document is signed or whether it is signed on all pages or whether the sign is matching with buyer/seller signature, are things we would want AI to do, so that the sales representative can spend more time understanding the customers’ needs. In short, AI assists humans and humans make better informed decisions.
Risk management use cases like AI-enabled fraud control, transaction scoring, and concurrent audit checklists, suggest that AI is becoming central to governance. How do you build trust in these AI systems internally and with regulators?
Interestingly, fraud control is an area where AI and ML has been prevalent for a long time. If you look at credit cards, most of the transaction monitoring is through AI, and has been in place for decades. So, the know-how of how to monitor these models are present.
Through an AI Annotation and hind sighting team, we have invested heavily in monitoring and maintenance of our modelling infrastructure. Our in-house document tampering model uses advanced computer vision algorithms and looks for multiple fraud indicators like Rasterization, non-standard margin, font and size from bank statements or salary slips, which are very difficult for humans to distinguish with naked eyes. These AI alerts go for human review and decision which makes for a more robust system than just AI or just humans.
Empower your business. Get practical tips, market insights, and growth strategies delivered to your inbox
By continuing you agree to our Privacy Policy & Terms & Conditions