How AI is reshaping the financial services industry

AI is reshaping financial services, and its usefulness is growing.

Artificial intelligence has moved from an emerging technology to an embedded part of the financial services industry. Firms are using AI across nearly every function—from portfolio construction and risk analysis to fraud detection, client servicing and back-office operations. What was considered experimental a few years ago is now part of a firm’s infrastructure.

The pace of change here can feel steep. But the direction is worth paying attention to, and is largely positive. AI is expanding the range of what's possible in financial services, giving firms and financial professionals better tools for handling complexity, spotting patterns and delivering semi-personalized guidance at scale—while also giving clients faster access to information and ideas.

But clients are drawing a clear line here. Investors are increasingly comfortable turning to AI for financial information—research, quick answers and a first pass at understanding an idea or solution. Actually handing over control is a different story: only 26% of investors say they'd let AI manage their investments. That gap is important. Clients want the speed and access that AI provides, but they still want a person making the calls.¹

It's important to remember what AI cannot do—no matter how fast, capable or well-integrated it becomes. That's where the financial professional's role continues to matter, even as AI's capabilities expand.

Machine learning and large language models are working together in new ways.

Two AI technologies are getting attention (and compliance scrutiny) in financial services: machine learning and large language models. They're often mentioned together, but they do different things.

Machine learning (ML) finds patterns in data. It's what powers the systems that flag unusual credit card or banking transactions, forecast market behavior and match clients to products. The more data it sees, the better it gets at recognizing patterns and the associated impact.

Large language models, or LLMs, work with language and text. They read, summarize, generate and translate—which is why they show up in tools that draft reports, answer client questions or pull key details out of long documents. (Any financial professional who's dropped a long legal document into an LLM knows how much time this can save.)

But useful isn't the same as reliable. LLMs are built to produce fluent, confident-sounding language—not to know when they're wrong. That's a meaningful risk in a field where the advice sounds credible whether or not it's accurate. Andrew Lo, a finance professor at MIT Sloan who has spent years studying how AI is used in financial advice, has cautioned that LLMs can generate persuasive financial advice that's also risky or simply incorrect—which is exactly why human oversight and thorough review must remain in the process.²

The combination of these two technologies is an attempt to address the reliability issue. Machine learning is increasingly being called upon to make LLMs more accurate and reliable. LLMs, in turn, are giving machine-learning models more accessible ways to interact with humans. Where they overlap is where the output gets interesting—increased fraud detection, jargon-free suggestions, portfolio insights delivered in conversational format or research assistants that can both surface data and describe plainly what it means.

The takeaway isn't the technology itself. It's that AI tools are becoming more integrated and more useful, but still require expert, human oversight. Taking on tasks that used to require multiple systems or manual work—including routine administrative work—is a trend that’s likely to continue.

Beyond ML and LLMs: a few other areas to watch.

Outside of machine learning and LLMs, there are other developments worth keeping an eye on.

Agentic AI. Rather than answering a single prompt, agentic systems can carry out multi-step tasks—researching a topic, pulling data from different sources, generating a draft and following up—with limited human input. Think: client onboarding, document processing and meeting preparation.

AI-native tools. These are built directly into customer relationship management (CRM) systems, financial planning software and portfolio tools to summarize meetings, prompt for a next-best action and semi-personalize client communications. Think: a virtual para-planner on your team.

Voice and conversation AI. Real-time virtual assistants can transcribe, summarize and even flag follow-up items during client conversations. Think: less physical note taking during phone calls or meetings.

None of these developments will fully replace the work financial professionals do. But it’s worth learning about, since they'll shape what tools become available and what clients come to expect.

AI is changing how markets move, how portfolios are built and how risk is managed.

The influence of AI across the broader financial system continues to grow. The technology is being used across market analysis, portfolio construction and risk management—often working behind the scenes on information the industry and clients rely on.

Market analysis and movement. Algorithmic trading powered by machine learning now accounts for roughly 80% of trading volume in U.S. equity markets.³ These systems can process news, price signals and data patterns at speeds no human can match—which has implications for pricing and volatility. The takeaway? Markets react quickly, and information gets priced in faster than it used to. The financial professional's role in helping clients plan thoughtfully, rather than react emotionally, matters more than ever.

Portfolio construction. AI screens large data sets, quickly identifies patterns across asset classes and stress-tests strategies against historical and simulated scenarios. This doesn't replace human judgment about a client's specific needs or risk tolerance, but it expands the range of scenarios that can be tested, giving financial professionals more nuance and context to bring into planning conversations.

Risk management. AI helps detect fraud, flag unusual account behavior and model complex planning scenarios that would be time consuming to run by hand. The industry is increasingly using it to spot risks and opportunities earlier and to act on them faster—which brings additional monitoring capacity to the systems that clients depend on.

None of these applications remove the need for a financial professional. Instead, they increase what they can offer, as more precise inputs and faster signals leave more time for the judgment, context and personal insight only a financial professional can bring.

Regulation, governance and transparency are catching up.

As AI use grows, so does the attention to how it's built, deployed and overseen. Regulators, industry groups and firms are all working to define what responsible AI use looks like—and the desired direction is toward more transparency, accountability and consistent standards.

Regulatory focus. Federal AI oversight remains a work in progress—the SEC withdrew an earlier proposed rule on predictive data analytics in mid-2025, while broader conflict-of-interest standards continue to apply.⁴

State-level activity. More than half of U.S. jurisdictions have now adopted the National Association of Insurance Commissioners' (NAIC) Model Bulletin on AI use by insurers, which sets expectations around governance, risk management and third-party oversight.⁵ More than 145 state AI bills were enacted in 2025 alone.⁶

Transparency for clients. As AI becomes more embedded in how financial services get delivered, clients are asking more questions about how their information is being used and how decisions are being made about their accounts. Being able to clearly articulate the firm's disclosure requirements—around both AI use and personally identifiable information (PII)—is becoming a bigger part of the client conversation.

Deploying AI in financial institutions comes with real costs and challenges.

Even with strong interest in AI, actual deployment isn't simple. Financial institutions are working through a set of practical realities that shape where AI gets used, the large investments around integration and how well it delivers.

Talent. Firms are competing for engineers, data scientists and AI governance professionals—many of whom are also being recruited by tech companies with different compensation structures. The challenges around recruiting and retaining talent can impact how firms prioritize AI initiatives.

Data quality. Many financial institutions have decades of client, transaction and market data spread across legacy systems that weren't built to work together. Cleaning, consolidating and preparing that data for AI use is often the biggest cost of any AI project—because the "garbage in, garbage out" principle applies to AI as much as it does to any other technology.

Integration with existing systems. Getting AI tools to work smoothly with the CRMs, planning software, portfolio platforms and compliance systems already in use is often a significant undertaking. Firms often find that the technology itself is the easier part; the integration work is where projects stall.

Cost and ROI. AI investments can be substantial, and not every use case delivers a return. Firms are prioritizing AI applications with a measurable impact—like efficiency gains in specific workflows, better fraud detection or improved client experience—rather than implementing every possible use case. This more disciplined approach is helping the industry mature past the initial adoption cycle.

Quantamental investing brings together data and human judgment.

One of the more interesting shifts in investing isn't purely about AI. It's about how AI and human judgment are being used together.

Quantamental investing—a blend of quantitative and fundamental approaches—is a good example of this in practice. Quantitative investing uses data, models and algorithms to identify patterns and opportunities across large numbers of securities. Fundamental investing relies on human research to assess a company's underlying business value, growth prospects and competitive position—the kinds of factors data alone can't fully capture. Historically, these have been treated as distinct traditions. Quantamental investing brings them together.

The term is used a few different ways across the industry, but the common thread is that AI supports human decision-making at multiple levels—both in analyzing broad market factors like valuation and momentum, and in evaluating individual stocks. MIT Sloan finance professor Andrew Lo has noted that large language models have created the opportunity to build a "powerful hybrid approach.” ⁷ An investment team may use AI to screen thousands of securities against dozens of factors, then apply human research to the shortlist. Or human analysts may generate a thesis and use AI tools to stress-test it against historical data, alternative scenarios and market signals. The AI adds speed and breadth. The human adds context, interpretation and the ability to weigh factors that don't show up cleanly in the data.

For financial professionals, quantamental investing is a useful reference point. It illustrates that the most effective use of AI in financial services isn't in replacement—it's in working together. AI can process information faster than any human can. But it can't replace the relationship, the reading of a client’s body language or the years of experience in interpreting how the strategies might play out in real life.

The financial professional's role isn't shrinking—it's changing.

As AI continues shaping financial services, only some of the changes will be visible to clients. Much of it will be embedded into platforms, workflows and decision trees that clients never see. Whether visible or not, it’s important to decide where and how to leverage AI in practice.

Clients still need someone who can listen, read the room, ask the right questions and interpret information in the context of real life. An LLM has never opened a personal retirement account, paid their own monthly bills, taken care of a family member, been laid off or blown out candles on a birthday cake.  AI can process, it can generate and it can suggest. What it can't do is sit across from a client during a hard decision, notice when something is unsaid or bring years of watching plans succeed and fail into a single moment of judgment.

Use AI to handle what machines do well—data processing, pattern recognition and routine work. Keep the relationship building and trust. That nuance is what turns a good plan into one a client actually follows through on.

Yes, AI is reshaping financial services. It’s also reminding the industry what financial professionals have always done best.