For years, artificial intelligence (AI) was evolving quietly in the background—mostly the stuff of research labs and tech companies. But that’s changing fast. AI is now moving into the spotlight, transforming everything from healthcare to entertainment. Even finance—an industry known for its cautious, heavily regulated approach—is starting to embrace it in a big way.
Financial institutions are now tapping into AI for tasks like market forecasting, fraud detection, personal finance tools, compliance work, and customer service. The shift didn’t happen overnight. Back in 2021, finance was still seen as lagging behind other sectors when it came to AI. Concerns around regulation, lack of infrastructure, and limited AI talent were big hurdles.
Then came 2023.
With the launch of large language models (LLMs) like ChatGPT and the rise of generative AI, interest in AI exploded across the business world. According to the International Monetary Fund, financial institutions are expected to double their AI spending by 2027. IDC predicts that, globally, companies will spend more than $500 billion on AI hardware and services by then. Clearly, the momentum is real.
And it makes sense: AI has the power to reduce human error, analyze massive datasets in seconds, and spot market trends faster than any human team could. But there are risks too. In the wrong hands—or without the right controls—it could lead to more sophisticated cybercrime or even trigger major financial disruptions. So while the adoption curve is steep, most companies are moving forward with caution.
To explore what’s actually happening on the ground, we spoke with three finance pros from Toptal:
-
Carlos Salas Najera, former Head of Equities at London & Capital
-
Arvind Kumar, a global consultant who’s worked with Goldman Sachs, KPMG, and EY
-
David Quinn, a wealth management consultant and firm owner
Here’s what they had to say about AI’s growing impact—and where it’s headed next.
So, How Is AI Actually Changing Finance?
In short: It’s helping finance teams move faster, work smarter, and serve customers better.
Carlos Salas points out that the industry really started to shift when major players like BlackRock went all in on AI. That move sent a strong message: evolve or fall behind. In 2023, BlackRock started using AI throughout its operations—to improve investments, deliver better client outcomes, and stay ahead of the curve.
AI can boost efficiency and support smarter decision-making, Salas says. Still, adoption hasn’t been universal. Some firms are held back by outdated systems, the complexity of integrating AI with existing models, or just plain skepticism.
But that’s changing. Tools like ChatGPT have helped normalize AI use in everyday life—and that’s bled into business, too.
“Reluctance has faded as more success stories emerge,” Salas says. “Plus, regulators are warming up to AI, even if the frameworks are still evolving. It’s giving companies more room to explore.”
4 Ways AI Is Already Making a Difference in Finance
1. Cracking Big Data
Finance generates a massive amount of data—but most of it never gets used. AI can change that.
Instead of just automating tasks, machine learning (ML) models learn from the data. They spot patterns and use them to make decisions or predictions. This is incredibly useful for things like reviewing loan applications.
Take JPMorganChase, for example. Not long ago, human analysts had to manually go through thousands of complex loan documents. It was slow, tedious, and prone to mistakes. So the bank built COiN, a platform that uses natural language processing (NLP) to scan and analyze contracts.
The result? What used to take 360,000 hours of human effort each year now takes seconds, with fewer errors.
Arvind Kumar says you don’t need to be a tech wizard to get value from AI. “Sometimes, I just describe a project I’m working on to ChatGPT or Gemini—what kind of client, what I’m trying to solve—and ask for ideas. You still need experience to separate the good advice from the bad, but it’s a great starting point.”
2. Smarter Portfolio Management
AI isn’t just good at sorting data—it’s also good at using it to make fast, informed investment decisions.
This isn’t new. Renaissance Technologies has been using advanced algorithms for years, earning its flagship Medallion fund an eye-popping 63.3% return annually from 1998 to 2018. What’s different now is that AI—especially LLMs—can go beyond traditional algorithms. They can interpret unstructured data (like transcripts, emails, or news articles) and adjust to new info in real time.
That’s exactly what BlackRock is doing now. Instead of using general-purpose models like ChatGPT, they’re building specialized LLMs trained on very specific investment data. The more focused the data, the less noise—and the better the predictions.
The goal? Read between the lines of earnings calls, anticipate market moves, and act faster than competitors. For traders and portfolio managers, that’s gold.
How Smaller Firms Can Benefit From AI Too
AI isn’t just for big banks with deep pockets. Smaller financial firms can also take advantage of it—without breaking the bank.
David Quinn, who used to lead wealth management at a fintech startup (which later got snapped up by a major U.S. bank), saw the potential early on. At the time, he was testing a custom algorithm to suggest trades that aligned better with clients’ portfolios, while also accounting for things like cash needs and upcoming withdrawals. “Each day, the algorithm would recommend trades. I’d look them over and either approve or reject them,” he says. “That feedback helped build a valuable dataset for training an AI that could eventually predict which trades a human would approve.”
The startup was acquired before Quinn could fully finish the project, but he didn’t stop there. Today, in his own wealth management practice, he’s still using similar technology to anticipate client cash flows. By analyzing data from custodial accounts, his system identifies repeating patterns—like monthly deposits or regular withdrawals—and uses that to train an AI to make predictions without needing human input.
The goal? Make smarter decisions and provide a smoother client experience, using AI thoughtfully and practically.
Using AI to Keep Up With Compliance
As financial regulations get more complex, staying compliant is more than just ticking boxes—it’s a constant challenge. AI can help make that job a little easier.
After HSBC was found to be one of 17 banks used to launder over $20 billion for organized crime, the company knew it needed a better way to spot suspicious transactions. It partnered with Ayasdi, an AI startup that uses advanced algorithms to flag unusual financial activity for review. Later, it also teamed up with Silent Eight to bring generative AI into the mix—automating time-consuming tasks like screening customers, monitoring transactions, and reviewing alerts.
The results have been promising. HSBC says AI has helped cut the number of investigations by 20%. That means less time spent chasing false alarms and more time focusing on real risks. Plus, with fewer errors and more streamlined processes, the whole compliance effort has become more efficient.
Better Customer Service With Chatbots
Chatbots have come a long way. Today’s AI-powered assistants can do much more than just answer FAQs—they’re helping banks deliver 24/7 service that actually feels helpful.
Take Erica, Bank of America’s virtual assistant. Launched back in 2018, Erica started with tools like predictive insights, fraud alerts, and personal budgeting help. But it didn’t stop there. By 2023, users were interacting with Erica 56 million times a month, using it to monitor subscriptions, check their credit scores, track refunds, and understand their spending habits.
And Bank of America isn’t alone. Other AI assistants like Citi Bot SG, NOMI from the Royal Bank of Canada, and Sandi from Santander are also stepping in to handle basic banking tasks—freeing up human agents for more complex issues.
Customers seem to like the shift, too. A Salesforce survey found that 81% of banking customers prefer trying to solve problems on their own first using tools like chatbots, before reaching out to a real person.
Staying Secure While Using AI
Security is everything in finance. These institutions deal with incredibly sensitive data—from personal info and account details to confidential business records. Bringing AI into the picture adds new challenges, because it usually means collecting and storing even more data.
That’s why strong security policies are a must. “Collecting more data makes our AI smarter, but it also increases our exposure,” Quinn says. “So protecting that data is critical.”
Regulations like GDPR in Europe, CCPA in California, and the EU’s new AI Act all set strict standards for how companies must handle data. And the consequences of ignoring them can be serious—fines can run up to millions of euros or dollars.
So finance companies need AI solutions that are secure, compliant, and transparent. That means encrypting data, controlling who has access, and building privacy protections right into their systems.
Clean Data = Smarter AI
AI needs good data to work well. If the information you feed into a model is messy, incomplete, or inaccurate, your results will be too.
That’s why data quality is such a big deal. “Even the best algorithm can’t do much with bad data,” says Quinn. “It’s not just about building a model once and calling it done. AI needs to keep learning over time—so we need to be clear on what data we’re collecting, how we’re cleaning it, and how we’re updating our systems.”
In other words, data isn’t a one-and-done job. It’s an ongoing commitment.
The AI Skills Gap
Even with great data and tools, AI doesn’t run itself. You still need people who know what they’re doing—and that’s another big hurdle.
A report from Rackspace Technologies found that two-thirds of IT leaders see a shortage of skilled AI talent as their biggest obstacle. The tech is evolving so fast that the supply of experienced professionals just can’t keep up with demand.
And it’s not just about hiring any AI expert. A study from Harvard Business School found that AI can improve performance by up to 40%—but only when it’s used by the right people on the right tasks. If you try to apply it in areas it’s not suited for, it can actually make things worse.
“AI takes tinkering,” says Kumar. “You might start with a big idea and think it’ll be a quick win, but the early results can be underwhelming. You have to stick with it. The breakthrough usually comes after a lot of trial and error.”