The AI-augmented investment professional

What happened

As outlined in our recent roadshows and presentations around artificial intelligence (AI), automation is revolutionizing the financial industry. The sector, characterized by its vast amounts of data and repetitive processes, is perfectly poised for the AI application era. As a result, industry experts see AI as one of the most interesting technologies in terms of growth for the financial industry vertical (market estimated at >$50bn by 2030, with an expected growth rate of 15-20% per annum).

Over the past two decades, basic systems have evolved into advanced applications. For instance, software solutions that assess a borrower's creditworthiness have progressed from simple decision trees to complex algorithms using thousands of variables, both numerical (e.g., income, credit score) and non-numerical (e.g., education, employment history).

However, not everything has worked well over this period. Ten years ago, robo-advisors were expected to capture a double-digit market share of the global asset management market, estimated at ~$100tn. Instead, they captured less than 1% due to their lack of personalization and bespoke advice (they execute predefined strategies without recommendations). The importance of a human touch was also probably underrated.

Generative AI offers new possibilities to overcome these issues. How are we, as asset managers, preparing for an AI-centric world? What are the latest AI developments in the financial world? The role of investment professionals is changing.

After introducing the topic with broad industry research, we shall use natural language processing (NLP) to dive deeper into the fintech firms and how they participate in the AI revolution. Finally, we will explain what we are doing at our level.

Impact on our Investment Case

Overview of AI in the fintech world

Advisors, analysts, and relationship managers must continuously adapt to the financial industry's technological developments. With the acceleration of AI, significant changes lie ahead for all.

Concretely, how will AI impact the financial industry in the near future? To answer this question, we can start by examining industry reports from McKinsey, Ernst & Young, and others, exploring online resources, or using tools like ChatGPT for the "statistically most probable answer."

From these sources, one will find out that typical applications of AI in finance include interacting with customers via chatbots, detecting frauds, reducing risks, offering personalized services, forecasting more accurately, and automating trading or bookkeeping. We published an article detailing these use cases and the applications of AI in finance. 

One could argue that such a broad overview is too generic and applicable to almost every industry. We need more specificity to identify the likely winners leveraging or benefiting from these drivers. Companies that innovate with AI-driven solutions should be tomorrow’s leaders; they are good candidates for investment.

A case study: getting industry insights with NLP

To dive at the micro level and help identify the individual players poised to be real AI innovators, we use natural language processing (NLP) and large language models (LLMs). These tools help distinguish companies that have genuinely embraced AI from those merely using it as a marketing tactic. If you are familiar with "greenwashing," prepare for "AI-washing."

We have selected 28 companies that can be considered “fintech leaders.” As a disclaimer, we are invested in roughly two-thirds of them. For the remaining companies, we consider most of them “legacy fintech players” with a growth profile that usually does not match our investment philosophy. These firms span all our fintech subsectors and universe, from next-generation financial services to financial software, payments, and blockchain.

During quarterly calls, C-level executives provide insights into their strategic developments, including AI ones. Our sample of 28 companies conducted 220 investor calls in the past two years, totaling over 4,000 pages of transcripts. Reading all this material would have taken a financial analyst approximately two weeks (at 40 pages per hour). Summarizing manually AI-related content within this material would significantly increase the estimated reading time as AI can perform this task in seconds. The analyst can obtain a summary of each company's developments and perform sentiment analysis.

The first task of the NLP analysis is to summarize the development of AI applications and the evolution of AI investments for our selected companies.

For instance, with Guidewire Software Inc, a leading insurance software developer, we quickly learned that the firm is testing AI to facilitate cloud migration, potentially boosting its recurring revenue. The company is also expanding its generative AI features to improve its underwriting processes and claim management. If this strategy is well executed, it would reduce operational burdens and costs for its clients, reinforcing its competitive position.

Conversely, Fidelity National Information Servcs Inc, a banking software and payment giant, has disclosed very few developments in AI. This is unsurprising, as the company is engaged in a restructuring. However, it raises questions about whether the firm is genuinely building a new business model or merely addressing past issues.

In this article, we shall not provide a detailed analysis of each company, which we keep for our research records. But, it is worth examining the essence of the summaries to identify AI trends within our selection.

First, it is reassuring to see that some important applications, such as enhanced fraud detection, risk management, customer service automation (e.g., chatbots), and personalized customer experiences, are quite consistent among our selection of companies, in line with the broad industry's research.

More interestingly, the monetization of AI projects in fintech is gaining traction. Although new AI applications may still take a few quarters to impact top lines significantly, companies already see significant benefits from automation and AI on expense management, enhancing profitability. The same companies expect revenue boosts through AI-powered solutions from enhanced user engagement and retention, service fee increases, and cross-selling opportunities.

Another noteworthy aspect is that, despite AI becoming a strategic priority, many fintech companies rely on external partners to expand their AI capabilities and market reach as they lack internal competencies.

“Generative AI” is catching up with “automation”

In addition to summarizing transcripts, we conduct a sentiment analysis on AI application development and investment progression for our sample of 28 companies. The analysis focused on 25 keywords (e.g., “generative AI”, “machine learning”, “AI budget”, etc.), partially selected using ChatGPT.

Without any surprise, the overall sentiment in the fintech industry regarding AI is positive. This is expected for an industry aiming to disrupt traditional finance through technology. However, for most companies, the subjectivity (i.e., the amount of personal opinion in the text) was medium to high. Many applications are still in development, implying that executives do not have all the answers yet. As adoption increases, we anticipate future company comments on AI will become more objective.

Some companies are positive outliers regarding the frequency of keywords. For example, in this category, we can find Intuit Inc, the accounting and tax software provider. The firm has made a strategic decision to incorporate generative AI across its product suite, thanks to an internally developed generative AI operating system. They implicitly recognize that failing to be among the first movers could result in replacement due to their activities' high automation potential.

Intuit, along with Thomson Reuters Corp, a leader in compliance and legal solutions, is one of the few well-established firms that are very positive about AI developments during conference calls. Other large companies are more muted on the topic. The payment industry also appears to be lagging, beyond using AI to reduce payment fraud. It is surprising that card networks and payment processing firms do not emphasize integrating generative AI more, given its potential to influence consumption behavior. Conversely, smaller fintech firms and challengers understand that offering innovative AI-based solutions can provide a competitive edge.

The success of ChatGPT has also shifted the industry's development priorities. Before Q1 2023, no firms publicly discussed “generative AI.” Back then, “automation” was the main focus of the companies. “Generative AI is now catching up, as illustrated in the following chart.

Generative AI will not only be used by accounting software (Xero Ltd, a challenger of Intuit, which is also developing its solution). Thomson Reuters has launched a product in legal counseling and intelligent document drafting. Intapp Inc, a leading platform to manage deals widely used by financial firms, has integrated generative AI to suggest commercial actions, from contacting a prospect to drafting emails. nCino Inc, another banking software provider aims to improve the loan cycle with its digital banking advisor. Zillow, the leading real estate marketplace in the U.S., is reinventing the home search market with generative AI through personalized results based on user behavior and preferences.

The possibilities of generative AI seem limitless and will transform everything.

Generative AI at AtonRâ

The industry's transformation also impacts us. Extracting intelligence from thousands of pages is a game-changer, as is coding sentiment analysis in minutes. It has never been easier to make the best use of technology.

Our use of AI goes beyond the research case described above. All our developers rely heavily on generative AI to code more quickly and accurately. In terms of communication, we also get ideas and help from such tools — starting with the creation of the image illustrating this article.

We are also testing how these tools can improve our investment process, especially in defining our strategies and building our investment universe. Stock selection and portfolio construction are likely the following significant areas to be revolutionized by generative AI. Why should an analyst spend days forecasting earnings when some models already achieve better results than the consensus? No matter how good an analyst is, the machine can quickly and more accurately process data.

As more daily tasks become automated, analysts will have more time to focus on complex and value-added activities. Among the analyst's future routine tasks, we foresee many activities related to artificial intelligence, such as interpreting AI-based results, training, and refining models with the proper data, and continuously learning new models. However, the only way for analysts to truly stand out will be through human intuition, ethics, and creativity in complex scenarios. Using judgment and experience to make strategic decisions should remain a competitive advantage, as well as maintaining close interactions with the client side of finance to explain the machine's decisions.

Our Takeaway

Given the current developments, we believe fully autonomous wealth management solutions will soon emerge. This means everyone could access a digital family office providing financial, investment, and tax advice and management based on individual life cycles.

Until autonomous wealth management becomes widespread, one thing is sure: Investment professionals who utilize AI will replace those who do not. This applies at both individual and entity levels. Entities that strategically embrace AI are more likely to thrive. This is why we closely monitor industry developments and do not hesitate to adjust our portfolios to favor companies prioritizing AI innovation.