We’re now two and a half years removed from the explosion of transformer-based LLMs and other generative artificial intelligence into the marketplace, and the ripple effects are only getting stronger and more wide-ranging across industries.
AI development and deployment is moving at a breakneck pace, but one thing is still as clear today as it was in November 2022: the biggest economic opportunity for this technology lies in automating knowledge work.
There are few major industries that are as dependent on knowledge work as financial services. Banking, wealth management, risk assessment, and financial analysis all require specific expertise, information processing capacity, and cognitive skills. To date, powering all this work has required people, and lots of them: more than seven million people in the U.S. alone. On top of their headcount, firms also spend over $100 billion annually on outsourced labor.
The combination of ATMs in the 1970s and online banking in the 1990s made the job of bank teller all but obsolete. But that was just one job function. What happens when every job function in finance can be automated in part or entirely? We’re about to find out:
- Citi research places banking, insurance, and capital markets among the top five industries whose jobs have a high potential for automation, with 54% of jobs in banking alone having a high potential for automation.
- Bloomberg Intelligence data indicates CIOs and CTOs at global banks expect to cut as many as 200,000 jobs globally over the next three to five years due in large part to AI automation.
The days of brute-force analytical work in finance are numbered.
The fourth wave
The emergence of AI in finance fits within the twenty-year growth of fintech, which has played out in three distinct but complementary waves. Each wave has been marked by the emergence of upstarts who have become essential fixtures:
- Wave 1: Infrastructure and payments (Adyen, Plaid, Square, Stripe)
- Wave 2: Consumer Finance (Affirm, Chime, Credit Karma, Robinhood)
- Wave 3: Enterprise Finance (Alloy, Gusto, Pilot, Ramp)
Felicis invested in all three of these previous waves (Ayden, Plaid, Credit Karma, Alloy, Gusto). With AI, we’re entering the beginning of the fourth wave, one dominated by autonomous finance. Autonomous finance is the use of AI to automate and optimize financial workflows , without or with minimal human intervention. So, what great companies are going to be built first in this new era, and what will they specialize in? We’re watching four spaces most closely.
1. AI financial analysts
This is the largest and most exciting opportunity at the intersection of financial services and AI. Not only are there more than 400,000 people in the U.S. employed as financial analysts, but the role is also one of the most outsourced functions worldwide.
Add that the median financial analyst in the U.S. makes roughly $100,000 annually (and the top 10th percentile makes $175,000+), and the picture of why this is such a high-value opportunity comes into focus. The potential extends across multiple functions, from investment analysis and portfolio management, to compliance monitoring and risk assessment.
We’re already seeing early applications of AI around automated financial statement analysis, anomaly detection in transaction data, and predictive modeling for market movements. Firms that master quantitative analysis automation will find applications far beyond financial services, expanding into any sector requiring complex numerical reasoning.
2. Onboarding and underwriting
KYC due diligence can consume up to 40% of onboarding time for new corporate finance clients. And not without good reason: sophisticated fraud techniques cost global banks $500 billion in 2023 alone, according to Nasdaq research. Yet despite dozens of software solutions tackling KYC and AML automation, the majority of enterprise spend (almost $30 billion annually) remains allocated to human analysts.
We're already seeing promising early applications in the market, particularly in due diligence automation, AML screening, and document verification. Beyond immediate use cases, there's also a compelling long-tail revenue opportunity in enterprise counterparty risk management that remains largely untapped. The use of AI to perpetrate fraud will also grow exponentially, creating a cat-and-mouse game between compliance teams and criminals that will make the deployment of next-generation detection and prevention tools all the more necessary.
The complexity of regulations across jurisdictions will complicate AI automation efforts, but this will create an enormous opportunity for AI-native solutions that can streamline onboarding while maintaining or improving compliance standards.
3. Voice-based services
With advances in AI voice models and transcription, previously human-only interactions can now be automated with uncanny quality and shocking speed. When you consider financial services firms currently account for up to 25% of global contact center spend, this is a massive opportunity for transformation.
Beyond simple call deflection, we see the potential for AI to handle complex scenarios like dispute resolution, account maintenance, and even personalized financial advice delivery.
Collections agencies represent another prime target. These operations handle tens of billions in collections annually through primarily manual processes that neither companies nor consumers enjoy.
4. Agentic payments
As AI agents begin to proliferate, handling everything from sales to customer service to shopping assistance, they'll need corresponding infrastructure to process payments. With publicly traded payment companies comprising over $1 trillion in market capitalization, this represents a massive opportunity for startups who move faster than the global incumbents.
The most interesting questions here involve authentication, authorization, and liability. How does a payment processor verify an agent is acting within its authorization parameters? Who bears responsibility for erroneous transactions, chargebacks, and fraud? The answers will likely be written over the next three to five years by some combination of global platforms and emergent AI-payment-focused companies.
While competing with incumbents is challenging, success in this category will produce some of the largest companies—think of the next VISA, but AI-native.
Founders welcome
If you're working on solutions to transform financial services with AI, we'd love to talk.
The complexity and regulatory hurdles that make this space challenging also create moats for the companies that get it right. We believe the winners in this space will have:
- Deep domain expertise in both AI and financial services
- Products that drive immediate, measurable ROI through labor replacement or augmentation
- Strong compliance and security foundations built in from day one
- Go-to-market strategies tailored to the unique selling dynamics of financial institutions
We're especially interested in companies tackling high-expense labor replacement or augmenting large pools of knowledge workers with specialized, quantitative AI solutions.
The best time to build in this space is now, when the technological capability is finally matching the scope of the opportunity. The wave of autonomous finance has begun.
Thanks to Eric Flaningam for your help with research for this post.
