logo
No shortcuts in trust: Airtel Payments Bank CIO on responsible AI and banking at scale

No shortcuts in trust: Airtel Payments Bank CIO on responsible AI and banking at scale

Time of Indiaa day ago

Focusing on partnershipsAggressive use of AI, promoting financial inclusivityDefining digital transformation clearly, without embracing every emerging technology
In the ever-shifting landscape of fintech, CIOs are no strangers to buzzwords like 'blockchain,' 'digital transformation,' and 'quantum computing.' Yet, as Pinak Chakraborty, CIO of
Airtel Payments Bank
, reflects, many of these terms remain aspirational rather than operational realities for most enterprises. 'We are far away from using blockchain right now,' Chakraborty admits candidly, adding that while blockchain offers immutability and decentralised trust, practical deployment requires extensive collaboration across multiple organisations — a challenge that cannot be underestimated.
For CIOs steering large, complex fintech ecosystems, this means adopting a pragmatic mindset. Instead of chasing every emerging technology, leaders must focus on solving real problems with proven solutions while keeping an eye on promising innovations. This approach includes recognising that digital transformation is more than a catchphrase — it's about embedding technology in processes and culture to deliver measurable impact. However, Chakraborty notes, 'I have closed my ears to the term 'digital transformation' because too often it's used without clarity.'
Collaboration and governance: The twin pillars
What stands out in Chakraborty's narrative is the emphasis on partnerships — with industry bodies, technology vendors, and regulators. 'No new technology can be done without collaboration,' he says. The fintech space, particularly in India, thrives on ecosystem-level trust and interoperability, exemplified by initiatives like UPI which required multiple stakeholders to work in unison.
Extended vendor contracts, integration complexities, and regulatory compliance are not merely operational hurdles but strategic considerations that CIOs must navigate deftly. Chakraborty explains how Airtel Payments Bank works closely with NPCI, UIDAI, and hyperscalers like AWS, Google Cloud, and Azure, alongside regulatory bodies preparing for laws like the DPDP Act. Such alliances enable the bank to deploy sensitive features such as facial recognition for customer onboarding while ensuring adherence to compliance and data privacy.
Artificial Intelligence (AI), too, is approached with rigorous governance. Chakraborty shares that the bank implements MLOps practices including model version control, continuous testing, and rollback capabilities to ensure auditability and reduce risks like model drift and bias. 'AI is aggressively used, but it does not replace humans. There is always a 'man in the middle',' he stresses, highlighting the balance between automation and human oversight crucial for trust in financial services.
Balancing innovation with inclusion and trust
For CIOs leading fintech innovation in India, the ultimate goal goes beyond technology adoption. It is about driving financial inclusion safely and sustainably. Chakraborty describes how biometric cues such as facial authentication combined with contextual fraud scoring using machine learning models have transformed risk management, enabling safer second banking accounts for underserved populations.
However, the journey is not without its challenges. The reliance on multiple data sources necessitates centralised data lakes with unified customer views, while decentralised AI models must be carefully orchestrated for effectiveness and governance. There is also the ever-present need to mitigate data bias and hallucination risks through fine-tuning and restricting AI use cases.
Despite these complexities, the vision remains clear: leveraging technology not as an end but as a means to deliver trust, transparency, and inclusion. As Chakraborty poignantly puts it, 'Technology partnerships, regulatory alignment, and ethical AI governance are long-lasting journeys. CIOs who embrace this holistic view will lead India's fintech future.'
As the digital financial ecosystem continues to evolve at a breakneck pace, CIOs must play a more decisive role not just in tech adoption but in shaping the very frameworks of trust, security, and long-term interoperability. For Chakraborty, the future isn't about adopting shiny new tools but ensuring the bedrock data, ethics, partnerships, and resilience is uncompromised.
'There's no shortcut to building for Bharat,' Chakraborty says. 'You need to think long-term, act collaboratively, and design for trust.'

Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Tier-III cities drive 62% of new insurance premiums: Report
Tier-III cities drive 62% of new insurance premiums: Report

Time of India

timean hour ago

  • Time of India

Tier-III cities drive 62% of new insurance premiums: Report

HighlightsA recent report from Probus, an InsurTech and insurance broking firm, reveals that Tier-III cities and smaller towns are now responsible for approximately 62 per cent of new insurance premiums in India for the current fiscal year. The motor insurance segment experienced significant growth of 25.6 per cent in fiscal year 2024-25, driven by increased awareness, digital access, and an expanded Point-of-Sale Person network. Fresh life insurance policies saw a growth exceeding 60 percent in fiscal year 2024-25 compared to the previous year, while Small and Medium Enterprise insurance emerged as the fastest-growing segment with 112 percent increase. A new report from Probus, an InsurTech and insurance broking firm, indicates that Tier-III cities and smaller towns are now significant contributors to India's new insurance business. The report states that these regions account for approximately 62 per cent of total new insurance premiums in the current fiscal year. This trend suggests a shift in how new insurance business is generated in India. Smaller towns, aided by digital payment infrastructure like UPI and increased internet access, are showing higher rates of digital adoption. Probus highlights that this is a consistent pattern, with the contribution from these regions rising from 58 per cent to 62 per cent over the past three fiscal years. For the fiscal year 2024-25, the motor insurance segment saw notable growth at 25.6 per cent. This was attributed to increased awareness, digital access, and an expanded Point-of-Sale Person (PoSP) network. The report also observes that local distribution methods, including app-based onboarding and regional language interfaces, have facilitated wider access to insurance products. What may have been a compliance-driven purchase is increasingly being viewed as a financial planning tool in these areas. Across insurance verticals, fresh life insurance policies showed growth exceeding 60 per cent in FY 2024-25 compared to FY 2023-24. Non-life insurance, encompassing motor, health, and SME segments, also experienced growth. SME insurance emerged as the fastest-growing segment in FY 2024-25, with a 112 per cent increase. Rakesh Goyal, director, Probus, commented on the findings, noting the resilience of the Indian economy and the role of Tier-II and Tier-III cities in this. He stated that Probus's approach aligns with the goal of 'Insuring India by 2047,' by providing digital tools and strengthening PoSP models to offer insurance solutions to consumers in these emerging markets.

Big news for Paytm, Google Pay, PhonePe users: Starting August 1, you won't be able to check..., even linked accounts, mobile numbers, autopay...
Big news for Paytm, Google Pay, PhonePe users: Starting August 1, you won't be able to check..., even linked accounts, mobile numbers, autopay...

India.com

time3 hours ago

  • India.com

Big news for Paytm, Google Pay, PhonePe users: Starting August 1, you won't be able to check..., even linked accounts, mobile numbers, autopay...

New Delhi: There is news for crores of users who use apps like Google Pay, PhonePe, Paytm. From August 1, National Payments Corporation of India (NPCI) is bringing new rules regarding the use of its Application Programming Interface (API). According to the report, this technical change will be implemented from August 1. According to a report by Financial Express, the purpose of bringing the new rule is to reduce the burden on the UPI system. The effect of this will be that there will be a limit on the balance you check from the UPI app. There will be a change in the auto payments you have set. Let us understand this in detail. Increased load on UPI system According to the report, the number of UPI payments is increasing day by day. Every month around 16 billion transactions are being processed. Due to such a large number of transactions, the load on the UPI system is increasing. In recent times, cases of misuse of the system by banks have come to the fore. Also, some technical weaknesses were also seen. To deal with these, some changes are being made from August 1. UPI service outage According to the report, in the last two-three months, there have been many incidents when UPI payment was down. It is said that on April 12, people were very upset due to the payment being down for 5 hours. This was the longest outage in three years. Because of UPI, many people have stopped keeping wallets or do not keep money in their wallets. In such a situation, if UPI goes down, people face a lot of trouble. The rules are being changed to deal with this situation. The report shows that even a slight interruption in UPI service causes trouble to millions of users. It is said that every second 7 thousand transactions are being processed through UPI. If UPI goes down for even a minute, then 4 lakh people are affected. At present, the number of people making UPI payments is said to be more than 40 crores. A recent investigation revealed that the main reason for UPI going down is the repeated API requests, due to which there is a load on the system. UPI was affected in March and April due to the high number of 'check transaction' API requests. Changes happening from August 1 NPCI has asked all banks and PSPs i.e. PhonePe, Paytm, Google Pay etc. payment service providers to control the 10 most used APIs by July 31. That is, if users check balance more often on UPI app, then now they will not be able to do so. According to the report, users will now be able to check balance 50 times a day on their app. How many accounts are linked to the mobile number, this too will not be seen more than 25 times a day. Apart from this, autopay payments like any SIP or Netflix membership payment will be done only in non-peak hours. Nonpeak hours include before 10 am, from 1 pm to 5 pm, and after 9:30 pm.

Why JAM needs re-coding for robust social protection
Why JAM needs re-coding for robust social protection

Hindustan Times

time4 hours ago

  • Hindustan Times

Why JAM needs re-coding for robust social protection

In her book Recoding America, Jennifer Pahlka recounts her experience co-leading the Employment Development Department (EDD) task force in California during Covid-19. Covid relief was delayed by weeks due to a backlog stemming from outdated technology and rigid rules. For example, minor name mismatches could trigger a manual review, significantly delaying relief payments. We are often surprised when we first read about such systemic failures in a technologically advanced nation. But this is a global problem, a combination of technological choices, government contracting modalities, and rigid administrative processes. At a very different point in our technological journey within government, we face similar challenges in India, as a recent piece about India's first Aadhaar recipient highlights; Ranjana Sonwane received her monthly entitlements only in April, after a wait of nine months, because of an error linking her bank account to her Aadhaar. We are fortunately at an earlier starting point regarding technology in the Indian State; many of these systems for enhancing frontline welfare delivery are yet to realise their full effectiveness. At the same time, full-stack solutions are already emerging and have even matured across many states in India. The struggles of even a single citizen should motivate us to envision a more robust social protection system, especially given the maturity of supporting digital public infrastructure, such as the linkages between Jan Dhan bank accounts through Aadhaar and UPI protocols on mobile devices (JAM trinity). It is helpful to start with the fundamentals. Why do eligible citizens sometimes get excluded from receiving their rightful entitlements? As Pahlka shows us, technology is only part of the problem. Instead, a core challenge lies in how technology is designed and deployed to identify eligibility. If we begin with the assumptions that most citizens are likely to cheat to receive entitlements they do not deserve, we end up building extremely rigid processes and benchmarks that raise the administrative burden on citizens in terms of documentation, proof, and visits to government offices, that eventually only impacts the poor, educationally and technologically unprepared citizen. As a policy implementation organisation, through our work at Indus Action, we have found that it takes nearly 10 high-touch transactions, including more than three visits to government offices, to secure access to entitlements like scholarships, pensions and maternity benefits. Solving this problem requires combining new technological solutions and reframing the core problem. Let's reframe first, and adapt our strategy to one where we are comfortable living with a small degree of inclusion errors. Some undeserving citizens might get access to entitlements, but lowering the bar for inclusion will help lakhs of marginalised citizens. Technology allows us to check these inclusion errors if we think innovatively. From over a decade of work with governments across India, we have learnt that there are three key challenges in welfare delivery that new systems and technology can solve. The first is discovery: How can governments discover the eligible citizen instead of the other way around? Second, documentation: How can the government design rules and leverage existing data to validate eligibility for low exclusion/inclusion errors? And third, delivery: How can governments fast-track applications upon eligibility check and redress grievances to ensure on-time and quality delivery of entitlements? Luckily, we have enough bright spots from within India for these three significant challenges. Regarding discovery, states now have better quality information to discover vulnerable citizens. States can estimate spatial and household-level vulnerability through family ID linkage across department databases and auto-validate eligible citizens into relevant schemes. For example, states such as Punjab, Rajasthan, Karnataka and Jammu and Kashmir are providing validation tokens to workers who have finished 50 plus days of MGNREGA building and other construction jobs, making them automatically eligible for construction workers' entitlements administered by the labour department. On the burden of documentation, apart from leveraging validations of other departments (for example, birth/death registries; school-going age/grade), the rules governing eligibility can be reviewed to move towards exclusion-based targeting or targeting vulnerable households instead of individuals. For instance, all informal workers can be broadly divided into three archetypes: Farmers, construction workers and other e-Shram registered informal workers like domestic workers and gig workers. Once the family ID is linked to the key occupational information (like KAALIA ID for farmers in Odisha or shramik ID with labour departments), eligibility matching can provide a report of schemes that could be compounded to the family during crucial life-cycle moments. Finally, on the burden of delivery, we need look no further than passport seva and other government to citizen (G2C) services that have been transformed with commitments to operational excellence metrics like TATs (turnaround times) and SLAs (service level agreements). The RTPS framework and Bihar model provide the ambition for delivering grievance redress services for major G2C entitlements in all states. With the latest developments in Gen AI, it is possible to reach India's digitally unprepared citizens through voice in their Indic language, eliminating text barriers. Whether we hear about one citizen not receiving access, or view this as a systemic failure, we have a critical opportunity for learning to iterate better delivery systems and move towards universal access. We have experienced Covid-19 and find ourselves today increasingly unsettled by climate-related shocks. The next time a public health or climate emergency occurs anywhere across India, we can, through our DPI infrastructure for states, immediately release social protection grants to an auto-validated and eligible set of vulnerable citizens. This future is almost here if we act on every error to iterate on our delivery systems. Tarun Cherukuri and Rahul Karnamadakala Sharma are with Indus Action, a policy implementation organization. The views expressed are personal.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into the world of global news and events? Download our app today from your preferred app store and start exploring.
app-storeplay-store