Priya Jaiswal is a prominent figure in the intersection of banking and global finance, known for her sharp market analysis and strategic oversight of international business trends. As we navigate the complex landscape of 2026, the boundaries between technology and traditional banking are blurring more rapidly than ever before. This conversation explores the shift toward artificial intelligence platforms acting as financial “harnesses,” the competitive response from legacy banks, and the escalating “dead sprint” against a new generation of automated fraud. We also discuss how the industry is moving from simple data analysis to real-world financial action, changing how consumers interact with their money on a daily basis.
How are major AI platforms evolving from simple conversational tools into what are now being characterized as financial “harnesses” for everyday transactions?
We are seeing a profound structural shift where platforms like OpenAI’s ChatGPT and Anthropic’s Claude are no longer just answering questions but are becoming direct monetization levers. These AI “harnesses” are beginning to issue their own cards and collect interchange fees directly from transactions embedded within their interfaces, effectively positioning themselves as the primary gateway to a consumer’s financial life. This move toward card issuance is a very straightforward first step in their roadmap, allowing these tech giants to capture revenue that traditionally belonged to banks. By integrating these functions, the AI becomes the active layer through which money moves, rather than just a place where information is sought. It represents a fundamental change where the platform managing your schedule and emails is now also the one managing your purchasing power and payment flows.
With the introduction of agentic experiences like Meta’s Muse, what strategies are traditional banks and fintechs employing to ensure they don’t lose the customer relationship entirely?
The banking industry is acutely aware of the threat, and CEOs like Fifth Third’s Tim Spence are pushing back by building “bank-grade” agentic experiences directly into their own applications. The prevailing strategy is to ensure that high-stakes financial conversations, such as those involving mortgages or long-term lending, happen “within the four walls” of the bank’s secure environment rather than on an external AI platform. Banks are working to launch their own version of agents that have access to much richer, proprietary data, allowing them to provide advice that is more precisely tailored to an individual’s specific financial history. While the early chatbots we saw back in 2020 were largely ineffective and eventually disappeared, this new wave of AI agents is far more sophisticated and capable of handling complex queries with a high degree of personalization. It is a race to prove that the bank’s own app can offer the same conversational ease as a tech platform but with significantly higher levels of information security.
Which sectors of the financial market are currently facing the highest level of risk from this AI-driven business disruption, and how quickly must they adapt?
According to market risk assessments, the most exposed institutions are those catering to households and small-to-midsize businesses, particularly between 2026 and 2031. This is largely because the financial products these banks offer—like standard savings accounts and basic loans—are high-volume, standardized, and easily replicable by an AI agent. The risk here is moderate but persistent, as the impact depends almost entirely on how fast consumers change their preferences toward AI-led service models. If a customer can rebalance their checking and savings accounts through a voice command on a general AI platform, they may never open their bank’s mobile app at all. To survive this, institutions must differentiate themselves by offering services that an external “harness” cannot easily mimic, moving away from commoditized products toward more integrated and value-added financial planning.
The financial cost of generative AI-enabled fraud is a growing concern, but why do some experts feel that the current estimates of $40 billion in losses might actually be too low?
The projection that generative AI could drive U.S. fraud losses to $40 billion by 2027 is increasingly viewed as a conservative estimate because it doesn’t fully account for the emergence of “agent swarms.” These are coordinated groups of AI agents that can attack a bank’s infrastructure with a speed and scale that is genuinely terrifying to security teams. We are no longer just dealing with a single deepfake or a clever phishing email; we are facing automated, multi-pronged assaults that can overwhelm traditional defenses in seconds. It has forced companies like Plaid into a “dead sprint,” using AI specifically to fight AI-driven fraud in an ongoing arms race. The sheer volume of these attacks means that the financial toll could easily exceed those initial projections as the technology becomes more accessible to bad actors.
Given that the industry is currently in a “dead sprint” against fraud, how are companies like Plaid leveraging their vast networks to create a more collaborative defense?
Fraud is inherently a collaborative game, and with Plaid connecting to approximately 12,000 institutions, there is a unique opportunity to build a massive, shared network of data to spot anomalies. We are seeing a major push to give these 12,000 partners better security tools and real-time recommendations to protect their customers from increasingly sophisticated threats. The industry is moving toward a model where security is no longer a competitive advantage but a collective necessity, requiring banks and fintechs to share intelligence on new attack vectors. By tapping into this massive network, firms can identify patterns of agent-led fraud much faster than any single institution could on its own. This collaborative approach is the only way to effectively manage what many are calling a very scary year for financial security.
What is your forecast for how the customer-bank relationship will look as AI agents become more deeply integrated into our daily lives?
I believe we are heading toward a fragmented but hybrid future where no single entity owns the entire customer relationship, but rather different platforms serve different needs. For routine, low-stakes functions like moving money between accounts or tracking spending, customers will gravitate toward the convenience of AI “harnesses” that are already part of their digital workflow. However, for “big-rock” financial decisions like securing a mortgage or managing a complex investment portfolio, consumers will still want to interact directly with their bank’s own platform for the peace of mind and deeper expertise it provides. It would be crazy to assume that no financial functions will move to AI platforms, but it is equally unlikely that banks will disappear entirely; instead, they will evolve into specialized hubs for high-trust transactions. The ultimate winners will be those who can seamlessly bridge the gap between the convenience of an AI agent and the specialized security of a traditional financial institution.
