Unlike companies that treat “fail fast” as a badge of honor, financial services firms live in a different world. Every new technology has to survive a gauntlet of compliance reviews, regulatory standards, and risk assessments before it even gets close to production.

Our team hosted the Financial Services Frontiers: Real-World GenAI & LLMs webinar. This was an opportunity for industry experts to share a detailed account of how firms are innovating with Generative AI, their concerns and priorities, and the value this technology brings to users. Jacob Kossoff, Data Science and Model Development Executive at Bank of America, and Stavros Zervoudakis, professor at NYU, offered grounded perspectives on how financial institutions can stay innovative and compliant as they adopt AI in key operational use cases.

Acknowledging that missteps in financial services can damage customer trust, the webinar highlighted how firms are learning to manage these risks responsibly. At the same time, those that move thoughtfully and quickly can capture market share and secure a tech advantage that’s difficult to close later.

For financial services use cases, GenAI has reached a point where it is both powerful and reasonably reliable. Regulations are becoming more defined, and clients now expect experiences that are not only seamless but also intuitively intelligent. Our CTO, Justice Erolin, set the stage by saying that “What we’re seeing right now are the foundations of GenAI. What we’ll see in the future are the skyscrapers built upon those foundations.” That vision is already taking shape. GenAI’s impact can be seen in both the back office and in the customer experience.

Why GenAI Is Gaining Ground in Finance

Generative AI is carving a place in finance because it fits the sector’s dual mandate of running operations more efficiently while delivering smarter, more personalized services. Many firms are already piloting tools that can read, summarize, and cross-reference internal documents in seconds. It gives compliance teams and relationship managers the context they need without navigating hundreds of pages.

One of our clients in dental care financing used a GenAI chatbot trained on their own contracts to answer customer questions in plain language. This chatbot reduced support call volumes and shortened sales cycles, and gave customers the clarity they needed to feel confident signing agreements. Other companies are deploying GenAI to help users, such as analysts, spot patterns in unstructured data like news reports, market commentary, or regulatory updates that typically take days to navigate.

Adoption is accelerating in areas where measurable impact shows up quickly. Kossoff noted during our recent webinar that “six months ago, many firms had no internal chatbots, and now multiple departments use them to query internal templates, documents, and standards.” That kind of adoption speed used to take years, and now it is happening in months.

Externally, finance is a natural fit for GenAI’s client-facing potential. Zervoudakis emphasized that “there’s no excuse for financial institutions to not provide hyper-personalized service now, (…) It has a multitude of benefits for the company, not only brand perception, but also in improving processes for all users, whether it is an employee or an outside customer.” 

What Zervoudakis suggests resonates with a client’s project. A client of ours in the wealth management field piloted a system that lets advisors query years of client history in natural language. This allows them to prepare highly personalized engagements in minutes instead of hours. This shift is focused on meeting evolving expectations of both service providers and their clients.

In the webinar, Stavros Zervoudakis reminded us that a decade ago, this level of service wasn’t possible. But now, with smaller models processing sensitive data internally and larger ones handling broader tasks, there’s little excuse for financial institutions not to offer hyper-personalized services.

That momentum is energizing, but it’s not everything. In finance, long-term success doesn’t come from moving fast. It comes from moving with conscientious readiness. And that means pairing innovation with the right foundations to make it sustainable in a highly regulated environment.

Laying the Groundwork for Responsible Adoption

Generative AI’s potential in finance is enormous, but so are its complexities. Regulation is critical in the conversation, but just as critical as getting the fundamentals right.

When asked about the biggest roadblocks to putting GenAI into real use in financial services, Zervoudakis pointed to one that comes up again and again: access to the right data. Even with solid setups like RAG pipelines or agentic AI processes, he noted, it all comes down to making sure the data is accurate, easy to reach, and pulled from a single, reliable source. This is especially true when it’s powering customer-facing tools.

If that foundation isn’t there, the AI can’t really be trusted to deliver useful results. But when the data side is well-handled, teams can rely on those outputs. That’s when systems become actually useful, because employees can actually use those insights to work faster, stay compliant, and improve service.

Kossoff explained that large U.S. banks have been concentrating their AI efforts internally to balance innovation with oversight. The priority, he said, is helping employees become more efficient, more effective, and more engaged. That focus is guiding how banks deploy tools like GitHub Copilot or AI-assisted drafting systems. This internal-first approach also sidesteps some of the readiness gaps smaller institutions face in technical capacity and compliance.

Additionally, Zervoudakis underscored the value of starting small. “You need to start small, and you need to take adoption into your solution process,” he said, adding that scaling only works when risk frameworks are solid. This aligns with guidance from the Financial Stability Board (FSB) and consulting firms that call for centralized governance, explainable AI, and audit trails, especially in sensitive areas like credit scoring or anti-money laundering, where decisions must be traceable to earn regulatory trust.

Responsible adoption is more than just covering governance. It’s about building the right foundation. Even the best risk frameworks can fail if the underlying data, pipelines, and systems aren’t built to support AI at scale.

Data, Pipelines, and Systems Readiness

GenAI in finance is only as strong as the data and infrastructure behind it. That means solving for quality, accessibility, and compliance in one motion to ensure outputs are both accurate and defensible.

Our CTO, Justice Erolin, addressed this challenge in the webinar by pointing to hallucinations as a core risk in LLMs, especially when systems lack contextual or historical grounding. Zervoudakis highlighted that even with robust tools like RAG and vector databases, hallucinations can persist unless subject matter resources and documentation are contained.

Kossoff emphasized that avoiding “hallucinations” and compliance missteps starts with architecture. “Banks have historically built strong risk management (…) and now they’ve adapted and right-sized it for AI, bringing in tech risk, third-party risk, legal, and privacy at a larger scale.” This often involves hybrid pipelines combining on-premises and cloud resources, retrieval-augmented generation with validation layers, and logging mechanisms that make regulatory audits straightforward.

Well-designed pipelines also enforce governance at every stage of the data flow by managing permissions, monitoring latency, and securing integration points. This enables models to operate within compliance and performance thresholds from input to output. At the same time, AI model outputs in finance must often meet strict ex-ante and ex-post checks. This means validating models before deployment and continuously monitoring them for drift, bias, and explainability. This is particularly critical in functions like credit scoring or anti–money laundering, where errors can have systemic implications.

Investment reflects the scale of the challenge. Bain reports that financial services firms are investing an average of $22.1 million in AI—well above other industries—and staffing around 270 full-time equivalents to handle integration, governance, and scaling. This reinforces a key truth, and it is that readiness is the core of making GenAI work at enterprise scale.

People and Skills Driving GenAI Success in this Sector

As a software development partner who has worked on more than 1400 projects for the past 15+ years, we know well that technology itself doesn’t transform organizations. People do. And in finance, some of the most telling voices belong to CFOs. A Salesforce study found that only 4% of CFOs remain skeptical about AI, a huge drop from just a few years ago. It’s a strong signal of how the finance sector is thinking about AI.

Still, skill gaps are one of the biggest hurdles. Bain’s survey shows that 70% of financial services firms face shortages, especially in technical and risk/compliance roles. In earlier findings, we noted that even among firms actively deploying GenAI, only a minority have cross-functional AI steering committees, leaving many without a clear decision-making structure for balancing innovation and oversight.

Kossoff’s view reflects this reality: “We want to make sure we create hours and capacity so people can take their vacation, meet customers’ needs better, and have the tools to do their job.” This same principle guided a fintech lending client to deploy a GenAI system for prospect qualification. Based on the customer’s goals and financials, it automated tailored Q&A about loan terms and underwriting criteria, so sales teams could focus on relationship building and deal closing.

This mindset encourages an environment where human expertise thrives alongside automation. Success depends on integrating AI literacy into the broader culture so staff across finance, compliance, and operations can question outputs, flag anomalies, and adapt workflows without over-reliance on technical teams.

The talent strategy for GenAI in finance needs three threads:

1. Upskill existing staff so they can collaborate with AI effectively.
2. Bring in external software development partners with expertise in both financial regulations and AI deployment.
3. Ensure strong strategic AI leadership to bridge compliance, technology, and business goals.

This is where seasoned nearshore software development firms like BairesDev play a decisive role. We bring not just technical execution, but also the knowledge, experience, and consultant perspective gained from delivering software development to over 100 financial and banking clients. That level of industry-specific proficiency is critical in a sector where regulatory, technical, and operational demands intersect.

Closing Thoughts

GenAI is reshaping how institutions operate and how clients are served. The firms seeing the most success are those pairing innovation with a disciplined approach to governance, data readiness, and talent strategy.

Adoption in finance will never be a race to move the fastest. It will be a race to move the smartest, with the right guardrails in place and the right people guiding the way. Now the question is, is your organization truly ready to move from promise to production? If it isn’t, we should talk.

Want to dive deeper? Watch the full webinar replay or connect with our team to see how GenAI can be adopted in your organization.