In his Meditationes Sacrae, Sir Francis Bacon wrote “nam ipsa scientia potestas est” — for knowledge itself is power.
In Leviathen, Bacon’s former secretary, Thomas Hobbes, shortened it to the oft quoted proverb ‘knowledge is power’. Modern usage is so ubiquitous as to be cliché. What is often overlooked is the deeper meaning that comes with context. From Meditationes Sacrae:
"They establish wider limits of God’s knowledge than of power, or rather of that part of God’s power (for knowledge itself is power) by which he knows, than by that by which he moves and acts."
As it is in Bacon’s meditations, so it is with AI in the future of finance.
For whilst it is one thing to have knowledge of AI — even to acknowledge the significant role it will play in the coming months and years — taking action in a way that builds competitive advantage is another thing altogether.
Who dares wins, indeed.
For more than a decade we’ve been told that “data is the new currency”. We know the financial sector utilises vast amounts of data. From financial modelling and customer analytics to risk management and fraud detection, big data drives decision making.
That said, there is an abundance of raw data. According to the latest estimates, 402.74 billion Gigabytes are created each day.. . How, then, can firms convert raw data into actionable knowledge? How can that data be leveraged to create a competitive advantage?
“We have plenty of data, but we need to transform it into a way that is used for analytics.” — Daniel Pinto, President & COO, JPMorgan Chase & Co
In this paper, we explore the potential for harnessing Al to convert raw data into actionable knowledge. We'll touch on current big picture trends first, then go deeper in two areas: Investment Research and Automation. Finally, we'll get more speculative, exploring potential future use cases should Al continue to advance at the rate we've seen in the past 22 months
As we look to trends in Al implementation, it is important to note that there remains a large gap between Al investment and revenue realised.
In Big Tech, the gap is estimated to be $600 billion over the past year 4. Key to turning the corner is focusing on problem solving rather than implementation for implementation's sake.
“Business leaders should take a problem-first approach, focusing on challenges rather than technique. Stop fixating on implementing specific types of AI (e.g., deep learning, GenAI) – instead, think how the right type of AI can be combined with automation and human input to create a systematic approach to problem solving.” — Shelby Austin, founder and CEO of Arteria AI Inc.
AI can process vast amounts of information quickly, but human intelligence remains critical for interpreting complex, nuanced financial decisions. On the client side, building relationships still requires human interaction, although that could change with a younger demographic and more advanced Large Language Models (LLMs). Human intelligence is also crucial in understanding factors like company culture or geopolitical risks.
AI tools can already track competitors’ strategies much faster than humans. Over the coming year, expect firms to integrate AI-driven competitive intelligence tools to gain an edge. Potential use cases include M&A, strategic investments, and product launches.
Clients are demanding more transparency and insight into their investments. AI-driven platforms can deliver on this by providing tailored, data-driven reports and market analysis. We’re all aware of the changing (or disappearing) face of the high-street bank and the move to online-only services. Whilst the traditional look and feel of the high street has changed as a result transformation and cost-cutting, Fintech challengers have added pressure on banks to adapt and adopt new ideas. The proliferation of personal finance apps - from wealth management with the likes of Nutmeg, Freetrade, or Swissquote, to money saving apps such as Plum, N26, or Yuh shows how far the retail banking world has evolved.
The idea that knowledge is power is more relevant than ever in the shifting landscape of finance. As Al and Machine Learning (ML) continue to evolve, it is action that will determine the winners and losers.
The future of finance will be defined by the rise of AI Automation, which will reshape everything from operational efficiency to client interactions and regulatory compliance.
The transition also presents challenges. For example, less human interaction with banking staff translates to less engaged customers, especially amongst older (and wealthier) clientele. Further, in its present form, credibility comes into question when considering the use of AI for deep analysis and assessment.
“I have the impression that credibility becomes more important in a lot of matters. If you just want to know ‘When was the inauguration of President XYZ?’ or ‘When did the Brexit vote happen?’, this actual information is very easy to get through AI. What you’re not getting is an assessment.” — Chief Economist, Investment Bank
Looking forward, there are speculative use cases with the potential to completely disrupt the financial industry, including analysis and assessment. (See Investment Research)
One of the most successful applications of AI to date has been updating the ‘digital core’. Whilst many financial institutions were still developing new systems using COBOL as late as 2006, since then most COBOL programming is purely to maintain existing applications. Meanwhile, millions of lines of outdated COBOL code handle most data and processing. Worse, it is often poorly documented and the number of mainframe experts and COBOL programme decreasing rapidly as they age out of the workforce.
Queue generative Al. Goldman Sachs reports positive use of Al to write code:
“Think of developers being able to auto-generate code from prompts. We’re starting to see that further boost productivity, with some developers saying they can write 20-40% of the code automatically in specific cases. Or, for example, they can create test cases for their code automatically. You create this kind of dualism, where you check the machine’s work, or machines must check human work. This kind of symbiosis is what is important.” — Marco Argenti, Chief Information Officer, Goldman Sachs
Argenti did not name the specific products and emphasised that it was still 'proof of concept' and not yet ready for production.
On the other hand, multinational IT services provider Accenture reports using Al tools to "rewrite millions of lines of our own COBOL code, surprisingly quickly and with great success".
Whilst high-net-worth clients often still prefer working with human advisors, many of today’s younger investors rely on robo-advisors to offer algorithm-based portfolio recommendations tailored to their client profiles. As young, tech-savvy, investors enter the market, robo-advisors could capture greater market share. To balance this trend, human advisors may need to reposition themselves as offering value-added services such as estate planning and holistic financial advice.
Although outside the scope of this brief paper, other AI Automation trends will affect the future of finance.
For example, J.P. Morgan Chase reports “[…] the value that we assign to our artificial intelligence use cases is around between $1 billion to $1.5 billion and is in the fields of customer personalization, trading, operational efficiencies, fraud manager, credit decisioning.” [emphasis added]
Other AI Automation use cases include Predictive Analysis, Regulatory Compliance AIs, and Algorithmic Trading.
Will Al and automation fully take over regulatory compliance functions? If so, how will that change the way financial firms manage risk?
Al is increasingly used to analyse historical data. Will Al's predictive capabilities in areas like market forecasting and risk management outperform humans in the next 12 months?
As of 2018, algorithmic trading accounted for roughly 75% of the overall trading volume in the U.S., European, and major Asian capital markets. 12 As Al-driven algorithmic trading evolves, could it affect market stability during periods of high volatility?
Investment research is one area where we see big things on the horizon. AI is increasingly used to analyse financial statements, market trends, and macroeconomic data at blinding speed. As AI tools become more central to research desks it will handle pattern recognition and forecasting, freeing human analysts to focus on high level insights and idea generation.
It is unlikely that AI will replace analysts any time soon. Instead, it will increasingly be used as a collaborative tool providing predictions and forecasts on a wide range of data. However, there are ethical concerns regarding bias in AI-driven research. AI is only as good as the data it is trained on. We’ve all heard of AI’s propensity for “hallucinations” and bias. With that in mind, there may be increased scrutiny on the ethical use of AI in investment research.
One potential solution is to train GenAI tools on past research reports of an analyst. It may then be able to create first draft reports that analysts can revise and update. Another way to increase accuracy is the use of ‘chain-of-thought’ prompts that mimic human reasoning.
In the future, AI’s pattern recognition capabilities may improve enough to identify warning signals of rare but significant market events. By definition, black swan events are difficult to predict using traditional methods, and under normal circumstances. In its present iterations, AI lacks contextual understanding of factors that lead to black swan events. However, as LLMs evolve and training data improves, AI may at least help mitigate the impact of such an event. This could be achieved through automated crisis-management processes, real-time data processing, and ML-enhanced risk management models.
Let's take a speculative leap into the future. Imagine jumping ahead five years. Al has continued to evolve rapidly, governments and regulatory bodies are on board, Big Tech and Finance continue to invest, and we now see Virtual Analyst Networks that work collectively across different sectors, industries, and markets.
What might this look like? First, let's start with a definition of swarm intelligence:
“Swarm intelligence is a form of collective learning and decision-making based on decentralized, self-organized systems. Natural examples are commonplace — flocks of birds and schools of fish act and react as groups, without instructions or direction from any single leader.”
Since it was first proposed in the 1980s, the idea of swarm intelligence has become interdisciplinary. There is ongoing research in the fields of economics, sociology, biology, and Al.
In the future of finance, swarm intelligence would provide an investor with bespoke recommendations based on personal preferences and investment goals. An investor could specify variables like sectors, industries, or markets, and the Swarm Intelligence Network would prioritise gathering the appropriate data.
Joe Investor wants to focus on EV technology. Since each AI “virtual analyst” in the swarm is an expert in a specific niche, the system would compile hyper-targeted insights so Joe could receive real-time updates on all things EV. The Swarm Intelligence System would pull from multiple AIs specialising in:
Based on this influx of data, the AI would recommend a buy alert for a specific EV startup about to secure a government contract, or a hedge against supply chain disruptions in rare-earth metals.
Now we get to the crux of the matter regarding AI’s role in the future of finance. AI is here to stay. Investment will continue, and positive outcome use cases will increase in frequency. It is easy to imagine that “this time is different”. Yet, as always, the more things change, the more they stay the same.
Whilst AI’s impact on finance may bring revolutionary technological changes, certain foundational aspects of finance remain constant.
The Importance of Trust and RelationshipsTrust is the bedrock of relationships between financial institutions and their clients. Algorithms may offer increased efficiency and precision. However, it is human relationships based on integrity and mutual respect that underpin the decision of where to place one’s wealth. Firms that continue to prioritise a balance between high-tech solutions and transparent, ethical interactions will retain (or gain) a competitive advantage.
The core purpose of regulation is to ensure stability, prevent systemic risks, and protect consumers. Put bluntly, that won’t change any time soon. It is possible that over the next five years, AI-driven RegTech will automate much of the process but humans will remain involved to ensure transparency and credibility.
There is no denying that AI can process data faster (and with more precision) than any human. However, high-stakes decisions involving unpredictable factors, ethical dilemmas, or moments of crisis still require human advisors and analysts. Until such time as AI can fully grasp nuanced context, it will take a back seat to humans in this respect.
