🏆
Global Knowledge Platform
Serving Researchers Since 2012

AI Adoption in Indian FinTech Firms: A Comparative Review of Digital Payment and Financial Service Platforms

DOI : 10.17577/IJERTCONV14IS110093
Download Full-Text PDF Cite this Publication

Text Only Version

AI Adoption in Indian FinTech Firms: A Comparative Review of Digital Payment and Financial Service Platforms

Brundha Anand

BMS College Of Commerce and Management Contact Email : brundha.bcom.2122006@bmsccm.ac.in

Theme 4: Business Innovation, Entrepreneurship & Digital Finance

Abstract – Artificial intelligence (AI) is moving from a back-office analytical tool to a core layer of Indias digital financial infrastructure, yet existing studies tend to examine a single firm or a single technology and rarely compare FinTech categories. This paper presents a comparative review of AI adoption across six Indian platforms: PhonePe, Paytm and Google Pay (consumer-facing payment apps) and Razorpay, Cashfree and PayU (merchant-facing payment gateways). Drawing only on secondary data (academic literature, regulator publications, official statistics, company disclosures and industry reports), the study examines AIs contribution to customer experience, operational efficiency, fraud management and financial inclusion. A transparent disclosure-intensity coding is used to compare platforms. Findings indicate that gateways disclose more AI activity in operational efficiency and fraud management, consumer apps in customer-facing automation, and that evidence on AI- enabled financial inclusion is the thinnest across all six. Data privacy, regulatory compliance and ethical AI emerge as shared constraints. The paper proposes a conceptual framework linking enablers, business-model context, AI capabilities, outcomes and moderating constraints, and sets out propositions and directions for future empirical research.

Keywords: AI; FinTech; Digital Payments; Financial Services; Customer Experience; Financial Inclusion; Digital Innovation; India.

  1. INTRODUCTION

    Indias Unified Payments Interface (UPI) has grown from 1.78 crore transactions in FY2016-17 to more than 24,162 crore in FY2025-26, with annual value of about Rs. 314 lakh crore (YourStory, 2026). At this scale, payment firms generate behavioural and transactional data that is well suited to machine learning, while fraud, failed payments and support load grow in step with volume. AI has therefore become a strategic capability for FinTech business models, not merely a cost-saving tool.

    Prior work on AI in finance has largely concentrated on banks, credit scoring or algorithmic trading, or treated FinTech as a homogeneous group (Cao, 2022; Ashta & Herrmann, 2021). Digital payment providers differ sharply in their business models: consumer apps compete for user attention and UPI volume, whereas gateways compete on merchant payment success and risk control. This paper asks whether those differences show up in how AI is adopted.

    The study addresses four research questions: RQ1 How is AI used across leading Indian consumer payment apps and merchant gateways? RQ2 How does AI contribute to customer experience, operational efficiency, fraud management and financial inclusion? RQ3 What data-privacy, regulatory and ethical challenges

    constrain adoption? RQ4 What conceptual framework explains adoption patterns across FinTech business models? The contribution is a cross-category comparison, a replicable coding of public disclosures, and a framework with testable propositions relevant to the conference theme of FinTech, digital payments and financial inclusion.

  2. Literature Review and Research Gap

    FinTech and business innovation. Gomber et al. (2018) interpret FinTech as a force of disruption that reshapes products, processes and business models, while Arner et al. (2016) trace its evolution after the global financial crisis. Digital payments are widely seen as the entry point for broader digital finance and inclusion (Ozili, 2018; Demirgüç-Kunt et al., 2022).

    AI in financial services. Cao (2022) surveys AI techniques and challenges in finance, covering fraud detection, risk analytics and customer service. Ashta and Herrmann (2021) discuss opportunities and risks of AI across banking, investment and microfinance, including opacity and bias. Explainability is a recurring concern (Arrieta et al., 2020), and Jobin et al. (2019) show convergence on principles such as transparency, fairness and accountability across AI ethics guidelines.

    Technology adoption lenses. Classic models, including the Technology Acceptance Model (Davis, 1989), diffusion of innovations (Rogers, 2003), the technologyorganisationenvironment framework (Tornatzky & Fleischer, 1990) and UTAUT (Venkatesh et al., 2003), explain adoption at individual or firm level. They say less about how a platforms business model shapes which AI capabilities it prioritises.

    Research gap. Three gaps follow. First, studies are often single-firm or technology-centric. Second, there is little comparison between digital payment providers and financial service platforms. Third, the joint contribution of AI to customer engagement, operational performance, financial inclusion and business innovation is insufficiently understood. This paper responds with a comparative, multi-outcome review of six platforms.

  3. RESEARCH METHODOLOGY

    The study is a structured narrative review (Snyder, 2019; Tranfield et al., 2003) based exclusively on secondary data: peer-reviewed literature, regulator publications (RBI, 2025), official UPI statistics reported in the press, company announcements and earnings commentary, and industry coverage published between 2024 and September 2026. Six platforms were purposively selected to cover three consumer apps and three gateways with national relevance.

    To compare platforms on a common basis, each was scored from 0 to 3 on four outcome domains using a disclosure-intensity rubric: 0 = no public evidence found in the sources reviewed; 1 = limited or indirect mention; 2 = documented deployment; 3 = documented deployment with a company-stated quantified result or target. Scores were assigned by the author from the sources in Table 1. They measure how much is publicly disclosed, not true adoption depth or effectiveness, and company-reported figures are unaudited. The scores are therefore indicative and should be read as a structured summary of evidence, not as a statistical test.

  4. MARKET CONTEXT: SCALE AND CONCENTRATION

    Figure 1 shows UPI volume rising from 4,597 crore transactions in FY2021-22 to 24,162 crore in FY2025- 26, while value reached about Rs. 314 lakh crore, growth of 30% in volume and 21% in value over the previous year (YourStory, 2026). Monthly volume touched 2,366 crore in July 2026. For the third-party app layer, NPCI data reported for October 2025 put PhonePe at 46.3% of volume, Google Pay at 35.2% and Paytm at

    7.5% (Inc42, 2025), as shown in Figure 2. This concentration means that AI investments by a few firms can affect a very large share of national payment traffic.

    Figure 1: UPI transaction volume and value, FY2021-22 to FY2025-26

    Source: Compiled by the author from NPCI data as reported by YourStory (2026) and Benzinga (2023). FY2023-24 and FY2024- 25 are rounded press-reported figures.

    Figure 2: UPI app share of transaction volume, October 2025

    Source: NPCI data as reported by Inc42 (2025). Others is the residual.

  5. COMPARATIVE REVIEW OF PLATFORMS

    1. Consumer-facing payment apps

      PhonePe. Its founder reported at Global Fintech Fest 2026 that automation resolves about 94% of customer- support cases without human intervention, with a 96% target, while handling close to 12 billion transactions a month; about 80% of fraud investigations are handled by automated systes (The Arabian Post, 2026). The firm has added natural-language payment search, AI-assisted merchant onboarding and SmartPages, which generates payment pages from plain-language, including vernacular, descriptions. Leadership stresses selective adoption: not every experience improves with AI.

      Paytm. Management credits AI-based fraud detection and collections for part of its FY2026 turnaround and says new investment will be directed to AI, using rented compute rather than its own data centre. It reports funnel conversion on agentic interfaces seven to eight times that of standard interfaces in a travel-booking pilot, and AI features on its Soundbox merchant devices (Outlook Business, 2026). The claims are management statements and are not independently audited.

      Google Pay. Public AI evidence relates mainly to scam protection. Google reports more than a million fraud warnings a week on Google Pay, on-device call-scam detection using Gemini Nano, and a pilot of screen-

      sharing scam alerts with Navi, Paytm and Google Pay (TechCrunch, 2025). Reach is limited: on-device detection initially works only on Pixel 9 and later devices.

    2. Merchant-facing payment gateways

      Razorpay. In August 2026 Razorpay announced Vulcan, described as a transformer-based foundation model for payments built with NVIDIA and AWS, trained on about 3 trillion data points from 4 billion payments. It claims 810% higher payment success, 8 times more international card fraud stopped and 5 times more fraudulent or disputed transactions identified without extra alerts (Amazon Web Services, 2026). Earlier, it acquired the AI start-up Thirdwatch for e-commerce fraud and return-to-origin risk.

      Cashfree. Its RiskShield system, launched in March 2024, applies machine learning across its merchant network and targets up to 40% lower fraudulent activity (IBS Intelligence, 2024). Its Agentic Payments Suite lets customers pay inside AI chat interfaces and uses voice AI to follow up on failed payments and abandoned carts; the claim that 8090% of failure reasons are solvable in one conversation is marketing, not an audited result (Cashfree, 2026).

      PayU. Announced in September 2026, Fraud Liability Protect uses machine-learning risk scoring on cross- border card payments to set the level of authentication and adds a layer in which PayU bears eligible fraud chargebacks. The company reports 45% better international card success rates, and one travel merchant reported a 14% rise in success and an 83% fall in fraud-to-sales ratio (The Arabian Post, 2026b).

      Table 1: Evidence summary used for coding

      Platform

      Type

      AI applications documented

      Company-stated result

      Source

      PhonePe

      Consumer app

      Support automation; automated fraud investigation; NL search; AI merchant tools; internal agents

      94% support cases automated; ~80% fraud investigations automated

      Arabian Post (2026)

      Paytm

      Consumer app

      Fraud detection; collections; merchant onboarding; Soundbox AI; agentic interface pilot

      78x funnel conversion in agentic pilot; FY26 profitability credited partly to AI

      Outlook Business (2026)

      Google Pay

      Consumer app

      Scam warnings; on-device call-scam detection; screen-share alerts

      1M+ fraud warnings per week

      TechCrunch (2025)

      Razorpay

      Gateway

      Vulcan payments foundation model; routing, fraud and risk scoring

      810% higher payment success; 5x8x more fraud caught

      AWS (2026)

      Cashfree

      Gateway

      RiskShield ML; agentic payments; voice AI recovery

      Up to 40% fraud reduction (target)

      IBS

      Intelligence (2024);

      Cashfree (2026)

      PayU

      Gateway

      ML risk-based authentication; fraud liability cover

      45% higher cross-border success; 83% lower fraud ratio (one merchant)

      Arabian Post (2026b)

  6. THEMATIC ANALYSIS

    Figure 3 presents the disclosure-intensity scores and Figure 4 averages them by category. Fraud and risk management is the most consistently documented domain: five of six platforms score 3 or 2, and all three gateways score 3. Operational efficiency separates the groups, with gateways averaging 2.67 against 1.67 for

    consumer apps, consistent with gateways selling payment success and risk control directly to merchants. Customer experience scores are equal on average (2.33), but through different mechanisms: conversational support and search in consumer apps, checkout and payment-recovery tools in gateways. Financial inclusion is the weakest domain, averaging 1.33 for consumer apps and 0.67 for gateways.

    Figure 3: Disclosure-intensity scores by platform and outcome domain (03)

    Source: Authors coding of public disclosures (Table 1). Dashed line separates consumer apps (above) from gateways (below). 0 = no evidence found; 3 = deployment with stated quantified result.

    Figure 4: Mean disclosure intensity by business-model category

    Source: Authors calculation from Figure 3 (n = 3 per group; descriptive only).

    Customer experience. Consumer apps use AI for support deflection, voice and vernacular interfaces and in- app search, and early agentic pilots point to conversion gains. Operational efficiency. Gateways use AI to improve routing and authorisation success, which converts directly into merchant revenue, while consumer apps report internal gains such as automated investigations and faster product review. Fraud management. Fraud is where AI is most mature, since labelled data and measurable loss make the business case clear. Financial inclusion. Vernacular and voice interfaces, such as PhonePes SmartPages, and merchant-facing insights on Paytms Soundbox suggest inclusion potential, but none of the sources reports inclusion outcomes such as first-time users, credit access for small merchants or reduced exclusion error. This is the largest evidence gap.

  7. CHALLENGES: PRIVACY, REGULATION AND ETHICAL AI

    Table 2 summarises the main constraints. The Digital Personal Data Protection Act, 2023 requires consent and purpose limitation for personal data, which affects model training on transaction histories. The Reserve Bank of Indias FREE-AI committee report (RBI, 2025) proposes a board-approved AI policy, data governance, model lifecycle controls, disclosure to customers, a way to contest AI-driven decisions and incident reporting for regulated entities; it is currently advisory but may be folded into Master Directions (Bar & Bench, 2025). Open questions include liability among institutions, developers and vendors, and the compliance burden on smaller firms.

    Table 2: Key challenges and mitigation themes

    Challenge

    Why it arises in payments

    Regulatory anchor

    Mitigation theme

    Data privacy

    Training and personalisation use sensitive transaction and device data

    DPDP Act, 2023

    Consent management; data minimisation; on- device inference

    Regulatory compliance

    AI decisions in a regulated payment environment; vendor and cloud dependence

    RBI FREE-AI report, 2025

    Board AI policy; model inventory; vendor contracts

    Algorithic bias and exclusion

    Fraud and risk models may wrongly block thin-file or low-income users

    FREE-AI fairness and customer-protection themes

    Bias testing; human appeal route

    Explainability and accountability

    Opaque models complicate disputes and audits

    FREE-AI; Arrieta et al. (2020)

    Explainable models for high-impact decisions; audit logs

    Evidence reliability

    Most outcome figures are company- reported

    None specific

    Independent audits; standard metrics

    An additional concern is the false-positive cost of fraud models. Aggressive blocking protects loss ratios but can exclude genuine users, which links fraud AI directly to the inclusion agenda and argues for reporting both fraud caught and legitimate transactions declined.

  8. PROPOSED CONCEPTUAL FRAMEWORK

    Figure 5 integrates the findings. Enablers (UPI-scale data, cloud and GPU compute, generative and foundation models) feed an AI capability layer of prediction, detection, conversation and autonomous action. The mix of capabilities deployed depends on business-model context, consumer app or gateway. Capabilities produce four outcomes, and outcome data feeds back to retrain models. Moderating constraints (privacy, regulation, ethical AI) shape how far and how fast firms move along a path from tool-level automation to platform-level foundation models and then agentic AI.

    Figure 5: Conceptual framework of AI adoption across FinTech business models

    Source: Developed by the author from the review.

    The framework yields four propositions for testing. P1: Merchant-facing gateways disclose and deploy AI more intensively in operational efficiency and fraud management than consumer apps. P2: Consumer apps prioritise customer-facing automation and conversational interfaces because competition is for user engagement at scale. P3: Documented AI-enabled financial inclusion outcomes lag behind other outcomes in both categories. P4: Stronger privacy and AI-governance requirements slow autonomous (agentic) deployment in high-impact decisions while raising investment in explainability and monitoring.

  9. IMPLICATIONS AND DIRECTIONS FOR FUTURE RESEARCH

    For managers, the review suggests matching AI investment to the revenue logic of the business model, and reporting outcomes on audited, comparable metrics. For regulators, standard disclosure of AI outcomes, including false-positive and exclusion rates, would improve comparability. For entrepreneurs, the gap in inclusion-focused AI, such as vernacular credit and merchant tools for micro-enterprises, is an opening for innovation.

    Future research could (i) test P1P4 using survey or panel data from FinTech managers; (ii) conduct firm- level case studies with primary interviews; (iii) measure inclusion outcomes with user-level data; (iv) examine agentic payments, including liability when an AI agent initiates a transaction; (v) extend the sample to lenders, wealth-tech and insurtech; and (vi) assess how the FREE-AI recommendations change adoption once codified.

    Limitations. The study relies on public, largely company-reported sources, a small purposive sample and author coding, and several 2026 announcements are recent and not yet independently evaluated.

  10. CONCLUSION

Across six leading Indian FinTech platforms, AI is most visibly embedded in fraud management, followed by operational efficiency and customer experience, and least visibly in financial inclusion. Gateways and consumer apps adopt AI differently, in line with their business models. Privacy, regulation and ethical AI are shared constraints that will shape the move to agentic payments. The proposed framework and propositions offer a base for empirical work on how AI creates value, and who it may leave behind, in Indias digital finance ecosystem.

REFERENCES

  1. Amazon Web Services. (2026, August 18). Razorpay launches Vulcan, Indias first AI payments foundation model fueled by NVIDIA and AWS [Press release]. https://press.aboutamazon.com/aws-international/2026/8/razorpay-launches-vulcan-indias-first-ai-payments-foundation-model-fueled-by-nvidia-and-aws-re- architecting-payments-for-a-350-bn-e-comm-future-by-2030

  2. Arner, D. W., Barberis, J., & Buckley, R. P. (2016). The evolution of FinTech: A new post-crisis paradigm? Georgetown Journal of International Law, 47(4), 12711319.

  3. Arrieta, A. B., Díaz-Rodríguez, N., Del Ser, J., et al. (2020). Explainable artificial intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information Fusion, 58, 82115.

  4. Ashta, A., & Herrmann, H. (2021). Artificial intelligence and fintech: An overview of opportunities and risks for banking, investments, and microfinance. Strategic Change, 30(3), 211222.

  5. Bar & Bench. (2025). RBIs FREE-AI framework: Navigating the legal and regulatory landscape. https://www.barandbench.com/view-point/rbis-free-ai- framework-navigating-the-legal-and-regulatory-landscape

  6. Benzinga India. (2023). UPI exudes digital payments dominance as transactions surge 82.2% in FY23, reveals NPCI. https://in.benzinga.com/content/31625438

  7. Cao, L. (2022). AI in finance: Challenges, techniques, and opportunities. ACM Computing Surveys, 55(3), Article 64.

  8. Cashfree Payments. (2026). Agentic payments suite [Company web page]. https://www.cashfree.com/agentic-payments/

  9. Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319340.

  10. Demirgüç-Kunt, A., Klapper, L., Singer, D., & Ansar, S. (2022). The Global Findex Database 2021. World Bank.

  11. Gomber, P., Kauffman, R. J., Parker, C., & Weber, B. W. (2018). On the Fintech revolution: Interpreting the forces of innovation, disruption, and transformation in financial services. Journal of Management Information Systems, 35(1), 220265.

  12. Government of India. (2023). The Digital Personal Data Protection Act, 2023.

  13. IBS Intelligence. (2024, March). Cashfree Payments launches RiskShield for real-time fraud management. https://ibsintelligence.com/?p=556831

  14. Inc42. (2025). UPI in October: PhonePe, Google Pays UPI lead softens amid record high volumes. https://inc42.com/buzz/upi-in-october-phonepe-google- pays-upi-lead-softens-amid-record-high-volumes/

  15. Jobin, A., Ienca, M., & Vayena, E. (2019). The global landscape of AI ethics guidelines. Nature Machine Intelligence, 1(9), 389399.

  16. Outlook Business. (2026). Paytm turned profitable and now it is going all in on AI. https://www.outlookbusiness.com/corporate/from-losses-to-profits-to-ai- paytm-writes-new-playbook-for-indian-fintech

  17. Ozili, P. K. (2018). Impact of digital finance on financial inclusion and stability. Borsa Istanbul Review, 18(4), 329340.

  18. Reserve Bank of India. (2025, August 13). Framework for Responsible and Ethical Enablement of Artificial Intelligence (FREE-AI): Report of the Committee. RBI.

  19. Rogers, E. M. (2003). Diffusion of innovations (5th ed.). Free Press.

  20. Snyder, H. (2019). Literature review as a research methodology: An overview and guidelines. Journal of Business Research, 104, 333339.

    /li>

  21. The Arabian Post. (2026). PhonePe scales AI across payments and operations. https://thearabianpost.com/phonepe-scales-ai-across-payments-and-operations/

  22. The Arabian Post. (2026b). PayU deploys AI fraud cover for international cards. https://thearabianpost.com/payu-deploys-ai-fraud-cover-for-international- cards/

  23. TechCrunch. (2025, November 20). Google steps up AI scam protection in India, but gaps remain. https://techcrunch.com/2025/11/20/google-steps-up-ai- scam-protection-in-india-but-gaps-remain

  24. Tornatzky, L. G., & Fleischer, M. (1990). The processes of technological innovation. Lexington Books.

  25. Tranfield, D., Denyer, D., & Smart, P. (2003). Towards a methodology for developing evidence-informed management knowledge by means of systematic review. British Journal of Management, 14(3), 207222.

  26. Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425478.

  27. YourStory. (2026, August). UPI completes 10 years, clocks nearly 13,000-fold rise in transaction volume. https://yourstory.com/2026/08/upi-completes-10- years-clocks-nearly-13000-fold-rise-transaction-volume

2 thoughts on “AI Adoption in Indian FinTech Firms: A Comparative Review of Digital Payment and Financial Service Platforms”

  1. Lucia Rossi says:

    Good comparative review across payment platforms. The differences in use cases for fraud detection versus customer engagement were well explained. Including regulatory compliance comparisons would add depth.

Comments are closed.