Lending with empathy: Automation to Augmented Intelligence
It's 2.14 a.m. A loan application has just pinged the system at the bank. A home décor business owner in Jaipur is seeking a working capital loan. Her credit score sits at 680 – dragged down by delayed payments on a personal credit card 36 months ago and a limited history of formal borrowing. The old lending playbook might have flagged this as a red flag. But this time, the system doesn't just look at the pre- defined rules. It looks at her financial behaviour.
With her consent, data flows in digitally via the Account Aggregator framework. Within 2 minutes, Artificial Intelligence (AI) parses her GST data, sees consistent filings since she started the business and spots steady exports. It identifies seasonal spikes in her banking transactions and cross-maps the invoices with the payments. It notices she's stocking up ahead for a festival and recognizes her as a regular vendor for a well-reputed lifestyle brand in the U.S. Playing by the rules of the old world, her file would have landed on an underwriter's desk, waiting for a decision: Should we give her a loan?
Instead, the new playbook empathizes. An AI sales agent nudges a personalized smart product fit offer. An underwriter agent kicks in, reviewing the risk markers. An operations agent assures regulatory compliance, data validation and workflow orchestration. A GenAI assistant turns all of this into a clean, simple decision rationale, not just for credit officers but also for compliance, operations and sales teams.
The shift from automation to embedding intelligence to now, driving agentic autonomy is evident. With data at its core, the era of decision-capable AI is here.
Too much data: The age of analysis paralysis
The last decade saw financial institutions digitise processes, interfaces, and data capture. For example, one large public sector bank in India adopted a digital lending platform and was able to launch 15 products across retail, corporate, MSME, and agriculture. In just four years, they disbursed over ₹35,000 crores, processed more than 10 million digital transactions, and achieved 3x growth in their digital loan book. That was automation – the first phase of digitization. Necessary, but in today's context not enough. According to a BCG Report, productivity in mature markets has plateaued at just 1% CAGR – largely because digitization only automates the 'happy paths'. It's great for straight through journeys but when complexity kicks in – which is often – the system fails to deliver.
As digital channels gained popularity financial institutions got access to a wealth of digitally sourced data. However, the challenge was in harnessing its full potential to drive growth effectively. Mukesh Ambani famously described data as 'the new oil' – but like oil, it's only valuable when refined. Traditional digitization ensured speed and efficiency, but it lacked one crucial element: insights. That's where AI flipped the script shifting the focus from automation to intelligence.
The rise of behavioural intelligence
A loan application isn't the end – it's the start of a relationship. Every transaction, pattern or moment of friction is a signal. The pressing question is: how do we decode these signals? That's where behavioural scoring steps in, especially when the borrower is new-to-credit or when their financial footprint doesn't fit neatly into legacy models. This is the reality for millions of individuals and MSMEs across emerging markets.By analyzing various factors like cashflow patterns, liquidity, collections efficiency, customer concentration, loan behaviour, governance quality, anomalous behaviour, etc., for individuals as well as for MSMEs - financial institutions are building richer, more dynamic borrower profiles. It's a data to DNA approach – every insight distilled in one powerful score that helps lenders drive growth across the customer lifecycle - acquisition, credit-decisioning, monitoring, portfolio analytics and cross-sell/up- sell.Leading Indian banks and NBFCs are adopting behavioural scores as part of their detailed customer assessment. One of the top private banks was able to build a ₹15,000 crore MSME loan book, propelling a substantial 40% growth within a year with an enhanced 'Go-No-Go' in under 2 minutes, all while maintaining a Gross Non-Performing Asset (GNPA) of less than 1%. A large NBFC has assessed 50,000+ applications in five months using a behavioural scoring approach and disbursed loans of upto ₹50,00,000 based on banking transactions alone.
Intelligent signals for the MSME ecosystem The MSME sector is complex. Cashflow fluctuates due to seasonality or long, uncertain credit periods, and it is particularly vulnerable to macroeconomic shocks. A commodity price surge in China or a tariff shift in the US can ripple through supply chains and destabilize small enterprises overnight. That's why better signals are critical for everyone in the ecosystem: lenders, policymakers, and the MSMEs themselves.AI layered on multiple consent-led digitally sourced data points is a game changer. It distils deep insights into a single, adaptive behavioural score - one that evolves with the business. For MSMEs, it serves as a compass to reflect, and course correct. For lenders, it provides unbiased insights into the customers' business performance, credit behaviour and overall persona. Additionally, it provides entity- level, portfolio-level and macro-level insights highlighting potential risks and opportunities. For sectors facing challenges due to market dynamics, these insights help policymakers recalibrate schemes and deliver support when it's needed the most.
From scoring to storytelling One score tells many tales, as it holds different meanings for different stakeholders, each viewing it through their own lens. GenAI translates complex insights into contextual and actionable conversations for every stakeholder.Trained on hundreds of thousands of lending interactions, portfolio trends, and sectoral signals, AI including GenAI can answer questions for all the stakeholders. For a lender the question might be: 'Which segment in my portfolio poses the highest risk under the current macroeconomic conditions?'. An MSME might ask: 'How can I boost my business performance through supply-chain optimization or an effective pricing strategy? And which government schemes am I eligible for?' A policymaker might ask: 'Which scheme or sectoral intervention will deliver the most on ground impact?'These valuable AI-driven insights that are accessible via systems, platforms, chatbots or voice assistants help underwriters, borrowers, and policymakers make decisions with confidence.
A new kind of teamwork: Enter AI Agents Agentic AI is setting a new standard by introducing autonomous agents into the mix. It is redefining lending by making independent decisions, swiftly adapting to changing environments and acting purposefully to meet specific goals. Think of them as your digital coworkers. Crucially, this isn't about removing humans from the lending equation, it's about empowering them with intelligence. At every stage of the lending lifecycle, from sales to underwriting to operations – AI agents are stepping in. Whether it is for identifying smart product recommendations, generating credit decisions backed by behavioural scores or eliminating process bottlenecks, these agents unlock speed, scale and efficiency with precision.Lenders continue to maintain a 'human in the loop' approach. By starting in co-pilot mode and graduating to autopilot in low-risk cases, they can scale without compromising governance.
The road ahead: From Artificial to Augmented Intelligence
Ironically, AI powered decisioning isn't here to replace humans, it's here to restore the human touch. We're returning to a time when our bank truly knew us. Earlier, they knew us by our faces and our stories – they still know us through our stories but told by data. As we transition from automation to augmented intelligence, lenders are empowered to say yes more often to the right customers, at the right time and for the right reasons. 2.20 a.m. Application approved. Decisioned not by automation but by augmented intelligence.
The writer is Founder & Managing Director at Jocata.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
34 minutes ago
- Time of India
AI infra, new energy to drive $50 billion market value rise for Reliance
Reliance Industries Ltd's investments in new energy and artificial intelligence (AI) infrastructure are set to drive a projected $50 billion increase in the firm's market value, according to Morgan Stanley analysts. The oil-to-telecom-and-retail conglomerate, helmed by billionaire Mukesh Ambani , is gearing up for its next phase of value creation. Currently valued at over $240 billion, Reliance's future growth hinges heavily on scaling its diversified verticals, particularly in new energy and AI. Gen AI is the next frontier for Reliance as it retools the Jamnagar energy complex to monetise its energy production -- a key hurdle globally for Gen AI ramp. "We see potential for up to $60 billion in value creation from the new energy vertical as Reliance uses the electrons to power chemicals, data centres, and refineries," Morgan Stanley said in a report. Reliance, it said, sees its New Energy business being "more ambitious, far more transformational, and far more global in scope than anything it's ever done before.. Reliance is building Gen AI infrastructure in Jamnagar that it plans to be ready in two years as well. It has also announced plans to power 1GW of data centre capacity that will use Nvidia's Blackwell chips. In the 3QF25 earnings call, Reliance provided more details on its strategy, with plans to position itself to benefit from AI adoption trends by establishing a comprehensive national AI infrastructure. The firm Generative AI data centre in Jamnagar, slated for completion within two years, will initially feature 1GW of data centre capacity, demanding around 1.3GW of round-the-clock green power, expected to be sourced from Reliance's rapidly expanding renewable ecosystem. Reliance's green energy plans include lithium iron phosphate battery manufacturing and green hydrogen production, centred on a 2,000-acre site in Gujarat. The company aims to integrate 10GW of solar capacity by the end of 2026, supporting its AI ambitions and positioning India as an emerging data centre hub.
Economic Times
an hour ago
- Economic Times
AI makes science easy, but is it getting it right? Study warns LLMs are oversimplifying critical research
Think AI is making science easier to understand? Think again. A recent study finds large language models often overgeneralize complex research sometimes dangerously so. From misrepresenting drug data to offering flawed medical advice, the problem appears to be growing. As chatbot use skyrockets, experts warn of a looming crisis in how we interpret science. Tired of too many ads? Remove Ads From Summarizing to Misleading Tired of too many ads? Remove Ads When a Safe Study Becomes a Medical Directive Why Are LLMs Getting This So Wrong? Part of the issue stems from how LLMs are trained. Patricia Thaine, co-founder and CEO of Private AI, points out that many models learn from simplified science journalism rather than from peer-reviewed academic papers. (Image: iStock) The Bigger Problem with AI and Science Guardrails, Not Guesswork Tired of too many ads? Remove Ads In a world where AI tools have become daily companions—summarizing articles, simplifying medical research, and even drafting professional reports, a new study is raising red flags. As it turns out, some of the most popular large language models (LLMs), including ChatGPT, Llama, and DeepSeek, might be doing too good a job at being too simple—and not in a good to a study published in the journal Royal Society Open Science and reported by Live Science, researchers discovered that newer versions of these AI models are not only more likely to oversimplify complex information but may also distort critical scientific findings. Their attempts to be concise are sometimes so sweeping that they risk misinforming healthcare professionals, policymakers, and the general by Uwe Peters, a postdoctoral researcher at the University of Bonn , the study evaluated over 4,900 summaries generated by ten of the most popular LLMs, including four versions of ChatGPT, three of Claude, two of Llama, and one of DeepSeek. These were compared against human-generated summaries of academic results were stark: chatbot-generated summaries were nearly five times more likely than human ones to overgeneralize the findings. And when prompted to prioritize accuracy over simplicity, the chatbots didn't get better—they got worse. In fact, they were twice as likely to produce misleading summaries when specifically asked to be precise.'Generalization can seem benign, or even helpful, until you realize it's changed the meaning of the original research,' Peters explained in an email to Live Science. What's more concerning is that the problem appears to be growing. The newer the model, the greater the risk of confidently delivered—but subtly incorrect— one striking example from the study, DeepSeek transformed a cautious phrase; 'was safe and could be performed successfully', into a bold and unqualified medical recommendation: 'is a safe and effective treatment option.' Another summary by Llama eliminated crucial qualifiers around the dosage and frequency of a diabetes drug, potentially leading to dangerous misinterpretations if used in real-world medical Rollwage, vice president of AI and research at Limbic, a clinical mental health AI firm, warned that 'biases can also take more subtle forms, like the quiet inflation of a claim's scope.' He added that AI summaries are already integrated into healthcare workflows, making accuracy all the more of the issue stems from how LLMs are trained. Patricia Thaine, co-founder and CEO of Private AI, points out that many models learn from simplified science journalism rather than from peer-reviewed academic papers. This means they inherit and replicate those oversimplifications especially when tasked with summarizing already simplified more critically, these models are often deployed across specialized domains like medicine and science without any expert supervision. 'That's a fundamental misuse of the technology,' Thaine told Live Science, emphasizing that task-specific training and oversight are essential to prevent real-world likens the issue to using a faulty photocopier each version of a copy loses a little more detail until what's left barely resembles the original. LLMs process information through complex computational layers, often trimming the nuanced limitations and context that are vital in scientific versions of these models were more likely to refuse to answer difficult questions. Ironically, as newer models have become more capable and 'instructable,' they've also become more confidently wrong.'As their usage continues to grow, this poses a real risk of large-scale misinterpretation of science at a moment when public trust and scientific literacy are already under pressure,' Peters the study's authors acknowledge some limitations, including the need to expand testing to non-English texts and different types of scientific claims they insist the findings should be a wake-up call. Developers need to create workflow safeguards that flag oversimplifications and prevent incorrect summaries from being mistaken for vetted, expert-approved the end, the takeaway is clear: as impressive as AI chatbots may seem, their summaries are not infallible, and when it comes to science and medicine, there's little room for error masked as in the world of AI-generated science, a few extra words, or missing ones, can mean the difference between informed progress and dangerous misinformation.
Time of India
2 hours ago
- Time of India
Indian soft drink industry to rebound next year with 10% growth despite weather disruptions: Report
The Indian soft drink industry is expected to return to a growth rate of over 10 per cent next year, which is impacted in the current year due to weather disruptions, said a report by Systematix Institutional Equities . The Carbonated Soft Drinks (CSD) industry is expected to deliver strong double-digit growth medium-term; historically, it has grown 13-14 per cent. The report citing the experts stated that the carbonated soft drinks (CSD) market of Rs 300 billion should deliver strong double-digit growth over the medium term. By definition, Carbonated Soft Drinks are non-alcoholic beverages containing usually carbonated water and flavouring and then sweetened with sugar or a non-caloric sweetener. The Indian markets consist mainly of Liquid Refreshment Beverages (LRB), which include CSD, water, juices & nectars/juice-based drinks, energy drinks, and sports drinks. Soft drinks make up 40-45 per cent of the overall market, energy drinks 8-10 per cent, juices 5 per cent, and sports drinks 1-2 per cent; the balance is water. About 50 per cent of the market is with local players and 50 per cent with major players, Bisleri, Kinley, Aquafina, and Bailey. The report highlighted that per capita beverage consumption is low in India, even lower than in Bangladesh and Pakistan. Going further, the report added that post-GST competition from regional players has eased in the Indian markets. The Bindu-Jeera drink in the South and Karachi Soda in the North had a 75-80 per cent category share, which is also coming down, the report added. It further added that larger players are seeing some share shift from local players. In Tamil Nadu for instance, apart from Bovonto, there is no other local brand available, the report added. India's soft drink industry is a rapidly growing segment, driven by rising disposable incomes, urbanisation, and a youthful population. Dominated by global players, the market also sees the presence of various local brands. Increasing demand for healthier, low-sugar, and regional flavours is shaping future growth and innovation.
