What will AI spell for the financial sector and careers there?
Markus Mäkelä
The financial services sector is among sectors where opportunities to use artificial intelligence tools abound. This is due to the nature of the most typical tasks in professionals’ daily and weekly workflows: they are often highly analytic, based on data that the company has access to, and often reducible to a quite deterministic algorithm – a sweet spot for AI use when partly automating high-risk or number-intense processes. And while generative AI normally handles numbers badly, such use cases often present large analytic-AI (“traditional AI”) opportunities in turn.
Critically, the best firms will manage AI’s introduction processes well – paying much attention to also the non-technical aspects of change management.
This includes succeeding with factors like their transformation’s strategy connection, organizational aspects of change management, project valuation and selection, early-begun collection of proprietary data, data governance, and risk management at large. In practice, all this is difficult to optimize and manage for larger firms, but for the successful ones, AI can improve their competitive position instead of merely savings costs from processes.
AI’s use cases and an impending industry shake-up
AI opportunities in finance fall into many categories. Eric Schmidt of Google fame and a former Goldman Sachs AI head just opined in the Wall Street Journal that beyond improving “antiquated and inefficient workflows in finance,” AI’s most transformative roles in finance involve “improving the core functions of markets themselves: pricing assets, measuring risk, detecting shifts in the economy, and managing volatility.”¹
Well, working with those tasks is part and parcel of the work of finance professionals, and I am therefore covering those applications of AI anyway as well.
I have curated an example list of valuable AI-use opportunities in some key financial services industries (below). Several use cases are of course valuable across the sector, be it information personalization, summarizing news, or alerting about news which are about a confluence of topics that are relevant exactly for a particular professional.
Across the financial services industries, AI will also be enabling new business models supported by innovative fintech AI – or other AI, say AI for sales. This promises to shake up parts of the sector either soon or within 3–10 years, giving chances for tech-savvy financial entrepreneurs for a fast market-share buildup and scaring incumbents alike. AI will also change organization structures, as discussed below.
Incumbent firms – including maybe all partners and CFOs now reading me – need to be thinking about how they may be outmaneuvered by the nimble new competitors. “Face it,” as long-term London Business School professor and visionary strategist Gary Hamel put it, “out there in some garage, an entrepreneur is forging a bullet with your company’s name on it.”²
They might also be outmaneuvered by comparatively nimble giants – say, a Europe-wide or global commercial bank such as the JPMorgan Chase of Jamie Dimon. Dimon is the world’s Timo Ritakallio, what comes to an early-adoption AI push in banking!
Proprietary models and agentic AI
Many or all firms will also need to tailor some of the AI models they use by training them with their proprietary data. While this can be very valuable in numerous applications – due again to the nature of tasks in the financial sectors many workflows – another side of the coin is that this can quickly amplify risks. Model-specific and sometimes situation-specific risk assessments will be necessary. (Expect regulators to weigh in, going forward.)
Agentic AI, which is maybe the hottest enterprise AI topic in 2025, refers to technologies that can execute a sequence of tasks and adapt to the user contextually without specific and intermediate instructions. Agents are a development that will bring into all this its own flair – that is, once it gains technical sufficiency for the workflow task in question.
Agents could in large parts automate a plethora of processes across large firms, from early fraud detection to treasury liquidity management and portfolio rebalancing, though these and similar ones will naturally require human-in-the-loop key-stage approvals for the foreseeable period. By their nature, AI agents may significantly speed up many broad-based cross-functional processes and make them HR-light – say, coordinating and smoothening end-to-end credit underwriting across risk, compliance, and front-office teams. Agentic AI may grow into a powerful set of technologies and be a big chapter of its own in the story of the financial world’s AI transformation.
Agentic AI, however, will also bring a whammy-sized risk cluster, which includes high-impact, low-probability risks, even for black-swan events. We don’t know much yet on agents’ effects on risk management processes, beyond that the demand for good ones will grow. But the benefits of agents would generally be great – assuming that control and compliance problems are solvable.
Impacts on organizations, employees, and careers
For young alumni, it is particularly important to learn about how AI will affect worklife in finance and carefully think about career choices accordingly. That is because AI is reducing job-position supply across professional services currently, including not just across financial services, but also e.g. in management consulting, especially in entry-level jobs. On a longer term, the reduction can be very large, and in finance, it is powerfully stepping onto the pedal.
But don’t forget the positive aspects. AI will create new tasks and roles as well, exciting change, and maybe a prolonged period where the design and rollout of enterprise AI tools will require expertise from people in the intersection of finance and AI – maybe from young experts.
Although these changes have started already, generally, a young professional could currently probably still do much if they are putting aside time to find out AI’s potential or soon-feasible use cases in their, and adjacent, lines of work, and then put in consistent work to learn some key uses of AI tools available.
Finding experienced mentors who know about finance and AI would be useful. Also, entrepreneurial opportunities might open up (though I recommend first getting some financial-sector employee experience). As one instance, some experts point to the innovation potential in the overlap between quantitative finance and AI³.
Having said that about entry-level positions, AI can begin to have large effects across an organization – e.g. by reducing also middle-management positions and even by buttressing much the feasibility of the more challenging organizational structures, such as an agile-team organization or a strict matrix organization with a “two solid lines” approach.
The large effects on organization of the business processes themselves underline the scale of the changes to come. Learning about AI’s opportunities and to use it effectively is a task relevant to all in finance; you need to be able to be a tight pair to AI in your work. One must learn to augment their efficacy by AI and keep learning more.
Markus Mäkelä is the President of AFA whose expertise includes business-process AI.
Selected AI use cases relevant for financial-sector industries⁴
General use opportunities
Pulling data, research, or forms; suggesting conclusions for the current work task.
Helping schedule meetings, summarizing online meetings, suggesting complete CRM system entries and meeting-prep steps from there, drafting follow-up emails.
Training young colleagues with AI tutors, which can, e.g., synthesize firm policy from several guideline documents for an individual, complex situation.
Generally, finding, sifting through, and synthesizing information quickly as well ask writing support for emails, slides, and other documents are important AI use cases, both for employees and (some) to offer to clients as a new digital service. AI will also come to enable personal and proprietary tool coding by many more employees than before. Agentic AI, for its part, will substantially broaden the set of general use cases by adding more execution-type workflows and AI autonomy. Moreover, AI’s ability to uncover subtle patterns in massive datasets can drastically improve financial decision-making; this phenomenon is often but not always about analytic, non-generative AI, and shows in multiple use cases listed.
Many of the listed benefits will be provided as embedded into your current software vendors’ products.
Several of the AI uses below have cross-industry applicability – say, between investment banking and investment companies for valuing a new strategy.
Asset and wealth management
Here, JPMorgan Chase has managed to boost sales, client numbers, and customer-service scalability. Underlying this was, among others, an ability to anticipate that certain clients would be asking certain questions during certain market and news developments, and pull relevant data for advisors in advance. Generally, gen AI has empowered JPMorgan’s professionals to focus better on high-value client work.
A word-for-word-tailored news report can be compiled automatically for all clients, judged based on their individual needs as estimated from their portfolio, interests and other needs.
Research
Pulling data, news and other research that is most precisely pertinent to a specific confluence of topic and situation-based interests for securities analysis; suggesting under-the-radar or weak-signal conclusions when contrasting the data with all recent news about the firm, its industry, or the broader economy.
Investment banking
Proposing targets based on news (e.g. when news details imply a company can suit a dormant or new client’s M&A-based growth strategy), drafting consequent client teasers, and creating, in effect, leads.
Identifying the research publications that most precisely comment on questions relevant for an individual case.
Quickly establishing from combined datasets that a project likely won’t have red flags. While this and several other listed examples need to be human-validated, finding red flags early can save great deals of time, which goes a long way to laying out the key rationale for many of the list’s examples. E.g. growth or customer views in key markets may rapidly prove worse than posited for preliminary valuations, or surprising contractual liabilities may surface.
Due diligence in general will be made faster despite that human verification must continue; also, “weak signals” can be spotted clearly easier.
First drafts of term sheets, letters-of-intent, etc., are produced fast. AI can extract terms for binding agreements or even write a fairly-complete first draft (before the necessary human work that would start from checking for comprehensiveness, intended focuses, and other quality).
AI can be used to suggest several strategy changes based on detailed company and external data, and if validated, they can support a noticeably higher valuation, should be acted upon by management, and can give a banker an edge in advising some clients.
More rarely, AI could pick up from company data something useful for brief prospectus mentions or e.g. point out a complex sales or cost synergy between several factors.
Intelligently proposing special or extraordinary valuation-relevant details from the depths of wide corporate data repositories, including details suggesting elevated risks.
Later, one may be able to interrogate an AI model for sensitivity analyses: how will cash flows change when changing this or that assumption? Or, what (else) must be true for such-and-such value scenario?
Private equity (some for fund management broadly)
Gathering and synthesizing intelligence for basis of documents for limited-partner relations, including fundraising. Automating parts of reporting to LPs.
Suggesting which foreign and possibly distant LPs may be relatively promising prospects to reach out to, based on a large multitude of data sources.
Lifting up opportunities to develop a strategy supporting a noticeably higher valuation target; AI tools with solid data access may be able to provide this. Examples how this can work include better product positioning to match customer needs with a comparatively inexpensive execution, and finding more of valuable cross-function synergies or sales synergies. See “Investment banking” above.
From combined pools of portfolio-company data, an AI could propose pooled synergies by the PE firm for just a certain type of fund portfolio, such as subprocesses of procurement, real estate, IT (e.g. AI!), or HR.
Entrepreneurs’ pitches to VCs should all be AI-augmented now, enabling you to demand a bit more of them. Have they, e.g. considered the strategic value of AI? In fact, you may need to see working products and revenue from them in your very first meeting, as explained below. Maybe most of all, they should display a general ability and attitude toward making broad use of AI for their business.
Early-stage companies are able to do more with AI, decreasing capital requirements by much. As a result, the American VC industry is doling out larger seed-round funding infusions (which entrepreneurs not rarely combine with bootstrapping), leaving less emphasis on later stages. There, industry data shows noticeable examples of lean growth. More and more American startups are even able to bootstrap their way through the early stages of business building, as IT management authority Thomas Davenport writes.
AI may possibly drive polarization in the VC industry toward mega-fund firms which can underwrite large-scale AI-based analysis and VC processes (and just possibly some target processes) and a number of niche-specialist analysts for these purposes.
While a private-equity investor’s assessment of target management remains key, what is needed from them is shifting – in not small parts. Also, your human-to-human relationship-building with them will naturally remain key, and AI can’t do all of that.
AI needs to be wisely embedded into target-company business processes, with good strategy-cognizant project prioritization and effective change management, creating at times very large and fast cost-saving opportunities.
Commercial banking and related retail services
Automating many parts of information retrieval and analysis for credit decision-making.
Fraud detection and prediction.
Compliance management at large.
Personalization of software-based services, whether internal or for clients, such as proposing a distant but well-suited, new investment prospect – a stock that well matches particular client preferences or portfolio optimization needs.
Automation of many directly customer-facing services in retail banking, such as having bots for servicing diverse customer information needs that actually work well.
¹ Eric Schmidt and Dimitris Tsementzis, AI can empower the financial industry, Opinion, Wall Street Journal (Nov 19, 2025).
² Gary Hamel, Bringing Silicon Valley inside, Harvard Business Review (September–October 1999).
³ Schmidt and Tsementzis, above. Such innovation, by the way, would also need to be supported via new, interdisciplinary academic teaching.
⁴ Related to JPMorgan Chase, see Alexander Saeedy, The AI strategy on the rise at JPMorgan, Wall Street Journal (Feb 27, 2025). Many of the VC-related observations benefited from Thomas H. Davenport and Erik A. Noyes, How generative AI is reshaping venture capital, Digital Article, HBR.org, Nov 18, 2025.