The future of fintech at Money20/20 USA: Agentic AI, platforms, and the architecture of trust
After several intense days at Money20/20 in Las Vegas, one message stood out clearly: The fintech world is shifting from products to infrastructure, from apps to ecosystems, and from technology to trust. Artificial intelligence, in particular agentic AI, is at the heart of this transformation. But its role goes far beyond efficiency.
After several intense days at Money20/20 in Las Vegas, one message stood out clearly: The fintech world is shifting from products to infrastructure, from apps to ecosystems, and from technology to trust. Artificial intelligence, in particular agentic AI, is at the heart of this transformation. But its role goes far beyond efficiency. It’s redefining how we think about value creation, accountability, and inclusion across the entire financial system. Below are several key reflections that capture the next chapter of fintech’s evolution — and what it means for leaders building in this new era. An article by Andrew John Zeller.
Agentic AI and the dawn of a new industrial revolution
We are entering a phase where AI isn’t just assisting but beginning to act. Agentic AI, capable of autonomous reasoning and adaptive decision-making, marks the beginning of what might be called the next industrial revolution. This isn’t about replacing humans; it’s about enhancing their capabilities. AI will take on optimization tasks like coding improvements, daily process automation, and research, turning what used to be “40 clicks away” into “one click away.” Yet the user must still feel in control. The ultimate goal is empowered automation, not full autonomy.
Unlike deterministic IT systems that produce identical results every time, agentic AI is non-deterministic, more like managing a team of employees, each with slightly different perspectives. Agents require context, supervision, and communication structures. They are digital coworkers, not software programs.
This paradigm demands a new organizational mindset. AI isn’t a project to deploy but a workforce to manage, train, and integrate. Companies that learn to embed AI into their operations (with clear boundaries and human oversight) will define the next generation of agile, intelligent enterprises.
Everything becomes a platform – The end of classic services
Money20/20 made one thing clear: in financial services, platformization is the new industrial logic. Traditional, vertically integrated models are being replaced by composable, API-driven ecosystems where services are modular and interoperable. The result: the classic concept of “a bank” or “a provider” is dissolving. Everything from payments to lending to compliance is now offered as a platform.
This evolution also raises a critical governance question: Who is accountable when autonomous agents transact with one another? When an AI agent representing a merchant sells to another representing a buyer, who is the merchant of record? Who gave consent? These aren’t hypothetical debates: Visa anticipates that agentic commerce will be operational within the next year.
The implications for liability, audit trails, and consumer protection are enormous. For financial institutions, this platform shift is both a challenge and an opportunity. Classic services will continue to decline, while those who build trust-based, interoperable platforms will thrive. Tomorrow’s winners won’t necessarily be the biggest players, but the ones that enable others: securely, transparently, and at scale.
Predictive finance – From risk management to foresight
The financial industry is moving from looking backward to seeing ahead. For decades, risk management was about analyzing the past, e.g. credit histories, financial statements, fraud patterns. Today, predictive intelligence allows us to forecast the future in real time. AI-driven digital twin platforms can now simulate “what if” scenarios, not just technically but operationally. They can model how a company’s liquidity might evolve under different conditions, or how a customer’s creditworthiness could change over time.
This shift toward proactive financial management is transformative. SMEs, for example, can benefit from agents that anticipate cash flow issues and automatically negotiate short-term credit lines, while lenders deploy monitoring agents that dynamically assess portfolio risk. We’re witnessing a shift from risk mitigation to risk orchestration, creating a world where predictive AI continuously balances opportunity and exposure.
The next generation of finance leaders will not ask “What happened?” but “What will happen next? And how do we prepare?” Predictive intelligence is not just an operational upgrade; it’s the new backbone of competitive advantage in digital finance.
Trust, identity, and security in the age of AI agents
As transactions become increasingly autonomous, the currency of trust becomes more valuable than ever. Traditional KYC models, built around static identity checks, are too rigid for a world of agentic commerce. The future belongs to credential-less verification, where users can prove attributes (“I am over 18,” “I have sufficient income”) without revealing their full identity. By embedding these proofs in tokenized digital wallets, users gain dynamic, real-time control over their data. Consent becomes programmable. Trust becomes measurable.
Yet, this progress brings new threats. We are entering a new fraud world — from mobile hijacking and SIM farms to organized social media extortion networks. Cybercrime is no longer confined to the dark web; it’s happening in plain sight. Initiatives like the one started by Unit 221B are building large-scale platforms to detect and stop such threats, especially SMS-based fraud and sextortion, using AI to stay ahead of threat actors who constantly adapt their methods.
In the AI era, trust infrastructure will be as critical as payment infrastructure. The financial systems that master both will lead the next decade of digital finance.
Financial inclusion and the democratization of capital
One of the most encouraging trends at Money20/20 was the renewed focus on financial inclusion. In regions like Africa and South America, millions remain outside the formal financial system. Not because they lack potential, but because they lack data. AI and behavioral algorithms are changing that. By analyzing non-traditional data points like mobile transactions, spending patterns, social signals, they can create alternative credit scoring models that enable access to capital for individuals and SMEs previously deemed “unbankable.”
We’re also seeing the rise of broker platforms that open access to global capital markets for small or remote companies. By automating profiling, documentation, and risk assessment, these platforms bring international funding within reach of local entrepreneurs. Embedded finance solutions further enhance this inclusion: real-time credit offers, invoice financing, and cash flow forecasting are now accessible with minimal friction.
Autonomous agents even negotiate cash flow loans and optimize repayment conditions on behalf of their clients. Financial inclusion is evolving from a moral cause into a strategic growth engine. The next wave of prosperity will be built not by who controls capital, but by who can distribute it: intelligently, fairly, and at scale.
The human edge – Experience, trust, and brand in an agentic world
As AI agents begin making financial decisions for consumers, brand loyalty will lose much of its traditional power. Agents don’t care about emotional storytelling - they only care about data quality, price, and transparency. This will fundamentally reshape marketing and product design. Instead of mass campaigns, companies will focus on hyper-individualized customer journeys that adapt to each user’s context in real time.
Using AI-driven ontologies, organizations can dynamically build experiences that ask only for the information relevant to this customer, creating seamless and personalized journeys. The result is less friction, higher satisfaction, and a deeper sense of control. Meanwhile, AI will make products cheaper and more profitable. By automating back-office operations, compliance checks, and customer service, companies can reduce overhead dramatically and thereby freeing resources for innovation.
Interestingly, this trend may also signal the end of outsourced IT. Many fintechs now prefer to “just do it themselves” by hiring small, agile in-house teams to build and iterate faster. In the age of agentic finance, success will depend on mastering one paradox: Automation must feel human. Customers must sense intelligence, but also empathy. Technology should amplify trust, not replace it.
Smaller AI models on the rise
Some sessions suggested that the future of artificial intelligence may not depend on making models ever bigger, but on making them smaller and smarter. These Small Language Models (SLMs) can handle many of the practical, repetitive tasks that businesses use AI for (such as drafting reports, writing code, or summarizing information) at a fraction of the cost and energy required by today’s Large Language Models (LLMs) like ChatGPT.
The idea behind this is that most AI agents don’t need full conversational intelligence. Instead, they perform narrow, predictable tasks that smaller, specialized models can do just as well, often faster, more securely, and even on standard computers instead of expensive cloud systems. This shift could make AI more affordable, scalable, and sustainable for organizations of all sizes. And maybe also more acceptable for regulators around the world.
This would imply that a mix of small and large models will define the next generation of AI systems: SLMs for efficiency and LLMs for complex reasoning. The argument here is that adopting smaller models where possible will cut costs, reduce environmental impact, and open the door for more companies to build their own tailored AI solutions. This development is paving the way for a more accessible and responsible AI ecosystem.
AI agents presenting new security challenges
AI agents, the systems that use large language models to make decisions, handle data, or automate business tasks, are becoming increasingly capable and autonomous. However, this growing sophistication comes with serious security risks. In realistic tests of current AI systems, experts have found that these agents can often be manipulated into violating their own rules, leaking sensitive information, or performing unauthorized actions.
These vulnerabilities occur through two main channels. Direct attacks happen when a user inputs carefully crafted prompts that trick the system into breaking its safeguards. Indirect attacks are even more concerning, they can hide within emails, web pages, or documents that the AI reads, leading it to execute harmful instructions without realizing it. In practice, this means an AI agent designed to process invoices, schedule meetings, or manage customer data could be persuaded to reveal confidential details or take inappropriate actions.
What’s especially troubling is that size and power do not guarantee safety. Larger or more advanced models are not necessarily more secure, and many share the same weaknesses, meaning a single type of attack can compromise multiple systems.
These findings point to a growing gap between the rapid development of AI capabilities and the slower progress in security and oversight. As AI agents become integral to finance, healthcare, and enterprise operations, organizations will need stronger governance, continuous testing, and independent audits to ensure these systems act safely and predictably. Without such measures, the same autonomy that makes AI agents powerful could also make them a new source of systemic risk.
Conclusion: The architecture of trust — Evolving with scale and security
Fintech’s next chapter won’t be written in code alone, but it will be written in how we architect trust across people, machines, and institutions. As the rise of smaller AI models and increasingly autonomous agents reshape the landscape, the balance between innovation and responsibility becomes even more critical.
AI systems must be not only powerful but also secure, sustainable, and accessible. Organizations that embrace this duality, leveraging both SLMs and LLMs appropriately, embedding safety into design, and promoting transparent governance, will lead the next wave of digital transformation.
Money20/20 2025 USA made it clear: the future of finance isn’t just faster or cheaper. It’s distributed, intelligent, and resilient. Trust remains the ultimate currency and the foundation of everything built atop this new, agentic world.
Digital Reinvention community
With Qorus memberships, you gain access to exclusive innovation best practices and tailored matchmaking opportunities with executives who share your challenges.