This Time Is Different: Why AI Is Unlike Any Wave I Have Seen In 40 Years Of Financial Services

Morris's perspective carries weight precisely because he has witnessed multiple technology cycles reshape finance. The internet, mobile banking, automation, and algorithmic trading all promised transformation. Yet each of these innovations ultimately enabled incumbents to strengthen their moats-they had existing customer bases, regulatory licenses, capital, and data that could be leveraged with new tools. AI, by contrast, attacks the fundamental cost structure and competitive differentiation of financial institutions.

The near-zero marginal cost proposition is particularly radical. In traditional finance, customer acquisition, onboarding, compliance, and servicing remain expensive operations even after decades of digitization. AI enables these functions to scale with minimal incremental cost, eliminating one of the key barriers that has protected incumbent market positions. A small team augmented by AI systems can potentially serve millions of customers with personalized service quality that was previously only available to the wealthy.

The economics of personalization shift dramatically with AI. Historically, personalized financial advice and custom products were luxury services available only to high-net-worth clients who could justify manual underwriting and customized structuring. The cost of human advisors, lawyers, and analysts meant that personalization only made sense for large account sizes. AI inverts this equation: algorithmic personalization scales infinitely without incremental cost, making it economically viable to offer tailored solutions to any customer, regardless of account size.

Consider mortgage lending as an example. Today's process involves standardized loan products, rigid underwriting criteria, and human review that limits the number of unique loan structures a lender can offer. An AI-driven system could potentially evaluate each applicant's cash flow patterns, employment history, family situation, and risk profile to offer dozens of customized loan structures-varying term length, payment schedules, rate resets, and default options-optimized to each borrower's circumstances. The cost to generate and service these variations would be negligible, yet the customer value would be substantial.

The financial services value chain has evolved slowly and deliberately over centuries. Banks collect deposits, assess creditworthiness, originate loans, and manage risk through regulatory capital buffers and diversification. Investment managers gather assets, construct portfolios, and charge fees for oversight. Insurance companies pool risk and use actuarial models to price premiums. Each layer of the chain exists because that layer performed a necessary function that others could not or would not perform. AI potentially disrupts this layered model by consolidating functions, automating judgment, and enabling new business models that don't require the infrastructure of traditional finance.

Consider the role of credit rating agencies and credit bureaus, which have guarded gatekeeping positions by controlling access to credit history data and judgments. An AI system trained on diverse data sources-transaction history, social network patterns, educational attainment, employment record, and many others-could produce superior credit assessments without relying on traditional bureaus. Over time, this could eliminate layers of the value chain that currently extract fees and wield outsized influence.

This is where the historical parallel to incumbent advantage actually turns pessimistic for large banks. Many incumbent financial institutions are exploring AI, but most implementations are aimed at improving existing operations rather than reimagining them. A bank might use AI to automate loan servicing, improve fraud detection, or personalize marketing offers within its existing product suite. These applications generate real value but fall far short of the paradigm shift Morris describes. They optimize the current value chain rather than reconstruct it.

The cannibalization problem is real and painful. If a major bank could build an AI-native lending platform that could serve customers at a fraction of current cost and with superior personalization, that platform would compete directly with the bank's current mortgage, auto lending, and small business lending divisions. Those divisions likely employ thousands of people, generate billions in revenue, and operate at healthy margins. Executives are naturally reluctant to cannibalize these franchises, especially when the upside is uncertain. This organizational inertia is not evil or irrational-it's a rational response to incentive structures and risk management imperatives. Yet it may leave incumbent institutions vulnerable to competitors willing to build AI-first platforms without legacy constraints.

Incumbent banks should not be counted out. They retain structural advantages: customer trust, regulatory licenses that are difficult to obtain, massive data archives that can train AI systems, and capital to invest in transformation. Yet these advantages only translate to competitive position if banks can overcome their organizational barriers to radical change. Some institutions may succeed by spinning out AI-native divisions with independent leadership, separate incentive structures, and explicit permission to cannibalize legacy revenue. Others may merge, consolidate, or deliberately shrink to reduce the burden of legacy operations. The worst outcome-and perhaps the most likely for some incumbents-is to be caught in the middle: investing substantially in AI while protecting legacy business models, resulting in the worst of both worlds.

The sources of competitive advantage in financial services are shifting. In the pre-AI era, advantages accrued to incumbents that controlled customer relationships (through branches and brand), data (through transaction histories), capital (to absorb losses), and regulatory licenses (difficult to obtain). In an AI-transformed era, advantages may shift to organizations that can generate and synthesize diverse datasets, build trust in AI-driven decision-making, move with speed, and pioneer new business models. These advantages are not inherently stable; they must be continuously renewed through investment, innovation, and organizational adaptation.

After 40 years in financial services, Morris has seen many waves. But the argument that AI is categorically different carries weight. Previous innovations enhanced efficiency or added capability; AI threatens to collapse cost structures, enable mass personalization, and undermine the structural advantages that have protected incumbent financial institutions. Organizations that navigate this transition thoughtfully-investing boldly in AI while ruthlessly pruning legacy operations-will likely thrive. Those that try to defend existing business models while dabbling in AI will likely suffer. And those that are built from scratch around AI capabilities may become the dominant financial platforms of the next era.