Gen AI in Financial Services: Technology’s Promise Meets the Human Element of Change

Written by Atul Kamra

Published on 02-20-2025

In my previous post, I discussed why legacy financial firms are well-positioned and open to technology innovations in the generative artificial intelligence (AI) era. These firms have the brand, the relationships, and the data, and they are attracting the talent. Now, how can they best use new technologies?

Let’s dig a little deeper.

AI and machine learning have been incorporated into offering financial services and finance for a while. What generative AI has done is transform the opportunity in terms of relatability and accessibility. Incumbents and startups alike are harnessing large language models' (LLM) capabilities for innovation and efficiency. They are leveraging trained models to perform tasks — and some reasoning — and the rapid improvements in AI training and inferencing costs continuously expand the scope of scalable use cases.

Don’t let me overstate it; models still require a lot of fine tuning.

Nevertheless, incumbent firms are balancing innovation with risk. Firms are starting with low-risk (i.e., low visibility), high-explainability use cases e.g., meeting summarization, enterprise search, data transformation, and deepfake identification. They are being cautious and experimental with high-stakes applications like credit decisions. And it is in these areas where startups are pushing the boundaries.

To be clear, generative AI in not solving for a new problem. The use case in summarizing meetings, making a payment, or making a loan is the use case. The problem is the same as before.

Generative AI provides a new, transformative solution that we did not imagine before. It has the power of converting static client interactions and workflows into dynamic, participative interactions.

Let’s take a familiar use case of chatbots. Incumbents that saw a chatbot as a terrible experience can now imagine a far more compelling, interactive user experience. Consider portfolio allocation and rebalancing. This was a one-sided process using pre-defined, limited inputs. With generative AI, it can become an interactive experience, where users actively engage with intelligent systems to participate in informed decisions.

As generative AI-powered agents or the bots get more sophisticated, there will be greater adoption — it will get the flywheel effect going for expansion into more use cases and user journeys.

Generative AI has changed the nature of competitive advantage. By allowing businesses and individuals to leverage data and automation, it will reshape how they differentiate themselves and build scale. Yet seizing (and sustaining) the advantage is not about the technology.

Technology is getting easier. What is getting harder is change.

To best prepare for generative AI it is necessary for financial services firms to rewire their talent and their working models.

• Larger incumbent firms and their teams generally operate in silos. To realize the efficiencies and enhance client experiences, team members must understand customer context and have domain knowledge across various interactions, functions and products.
• As companies venture into new use cases, there are many more ethical, regulatory, compliance and security challenges to overcome. While costs continuously come down for models and computing, compliance and security costs are a wild card in the medium term.
• Leaders and team members will need business training to work alongside AI. Without it, risk aversion will creep in to slow improvement and innovation.

Possibly the hardest to change will be incentives. Time (speed), continuous improvement and collaboration will be key performance factors. Being all-in on AI, companies need to measure and reward all three.

They need to measure for time (e.g., speed-to-scale a use case) and streamline their decision-making, reduce handoffs, and increase throughput in their service and product delivery. For this to be effective, business leaders and technical teams must collaborate.

Financial institutions are moving out of the winter of skeptical AI and technology adoption. They're proactive and seeing growth, leveraging existing strengths, and balancing innovation with risk management. The hard work will be adoption and change management.

The question isn’t whether AI will transform the industry — it’s who will execute best.