Ask a typical financial services company if it’s prepared to leverage AI, and you’ll likely hear a confident “yes.” In fact, a majority of businesses in this sector say they’re already using AI in production.
But if you dig deeper into how businesses in this industry are actually approaching AI deployments — if you ask questions like how they are governing their data, how they are ensuring data quality, and how easily are they connecting AI tools directly to data platforms — you’ll soon realize that claims about AI adoption in financial services don’t always align with reality.
That’s what I’ve long sensed when I assist financial services companies in AI adoption and acceleration. And now, thanks to a survey that Indicium recently released covering AI implementation in this industry, I have the data to prove that my anecdotal impressions are on the money.
What financial services companies say about the state of AI adoption
Financial services businesses recognize the tremendous value that AI technology has to offer by boosting their efficiency, accelerating processes, and helping them scale operations. According to the Indicium survey, 66.7% of financial services businesses currently have AI-driven solutions in production, and another 31.5% have deployed AI pilot projects or experiments. Not a single company reported having no plans at present to adopt AI.
Viewed from this perspective, it would appear that AI is rapidly becoming a cornerstone technology for financial services. Specifically, the survey reveals that financial services businesses are leveraging AI to support use cases like fraud detection, decision-making, and back-office process automation.
Lingering AI adoption barriers
However, when you assess the state of AI adoption in financial services based on how well AI systems actually connect to the data that drives them, a cloudier picture emerges.
- 40% of survey respondents indicated their AI initiatives are only “partially integrated” with their data infrastructure, suggesting that it’s tough to get data where it needs to be to drive effective AI-based automation.
- 40.6% said that, prior to undertaking data modernization initiatives, they were only “somewhat prepared” to use their data in AI tools and applications.
- 71.5% reported that preparing data for use in AI tools and apps was a major goal for data modernization projects, making this the top reason for modernizing data infrastructure.
Also revealing is the finding that, when financial services companies were asked which types of data they are using to power AI models, the most common response by far was structured data, such as information stored in databases, which 80.6% of companies are applying to AI implementations. In contrast, only 33.9% of respondents said they are using unstructured data, such as text documents and images, to power AI models.
In a way, this is unsurprising. It’s usually easier to work with structured data, given the fact that it is typically organized in a coherent, consolidated way — whereas unstructured data more often sprawls across multiple systems, making it difficult to capture comprehensively and feed into AI systems.
It’s notable that only a minority of financial services companies are prepared to leverage unstructured data as part of AI initiatives because part of the power of AI technology, especially generative AI and agentic AI, is its ability to leverage data of all types to power sophisticated decision-making, trend detection, and content generation. When businesses limit themselves to structured data sources to drive AI models, they are also severely limiting the effectiveness and flexibility of their AI solutions.
Accelerating AI implementation for financial services
For financial services companies interested in taking full advantage of AI, the takeaway is clear: Rolling out AI systems is one thing and connecting those systems to the data necessary to maximize their effectiveness is quite another — and it’s where businesses in this industry appear to be falling short.
Fortunately, this is a challenge that financial services companies can solve. The key is to modernize their data platforms and boost data governance and data quality by applying an effective data transformation framework. Equally important is establishing a robust, comprehensive approach to data operations, meaning the set of processes an organization uses to manage data.
When businesses embrace these solutions, they can more easily achieve high levels of data quality and robust governance, which in turn improves the performance of AI tools, since better data going into AI models translates to better model output.
In addition, enhanced data governance practices help to centralize all data around a single platform with built-in quality and governance controls. This also makes it easier to connect AI tools to high-quality data and to ensure the business is able to draw fully on all of its data assets — structured as well as unstructured — to drive AI initiatives.
Talking about data modernization may not be as exciting as talking about all of the interesting use cases for AI in the financial services space. But it’s clear that having this conversation is critical for businesses that are serious about using AI to full effect because without the right data — high-quality data and effective data governance controls — even the most cutting-edge AI solutions are of little value.
The article was written by Matheus Dellagnelo, who is the chief executive officer of Indicium, an AI and data consultancy based in New York City.
