How Can Financial Services Organizations Build AI-Ready Data Foundations?

Financial Services Leads the Way on AI

The Snowflake research found that financial institutions (FIs) are embracing AI to a greater degree and are spending more on the necessary data warehousing than other industries.

“Given both competitive pressures and the capacity for generative AI to improve digital customer experience and target compelling offers tailored to consumer need, the industry is likely to remain a leader in gen AI adoption,” the study’s authors write.

Early AI adopters among FIs are also far more likely to use the technology to improve financial performance of their firms (43%) than other organizations (30%). It’s no surprise, then, that cybersecurity (70%) and customer service (63%), two areas of any financial services business with high impact on financial performance, are among the most common uses cases for AI adopters in the industry.

UP NEXT: Five techniques for AI abuse and filter bypassing.

Using AI Tools to Make Data AI-Ready

Financial institutions working to make their data AI-ready must consider its format and whether it’s structured or unstructured.

Fortunately, generative AI is increasingly affording new experiences using unstructured data. For instance, the technology can simplify the scanning of thousands of PDFs to answer user queries, previously a cumbersome, error-prone activity.

Choosing the right generative AI tools to ensure smooth ingestion of unstructured data remains important, allowing a search experience to be layered on top — no data science background necessary, Avanes says.

Generative AI can also help organizations with their data governance challenges by automatically applying and enforcing access policies once data has been classified and, in some cases, handling tagging itself.

“You can find sensitive data and then basically instruct the enforcement of policies without needing to do so again or writing a ton of code yourself,” he says.

Click the banner below to refine your AI strategy with these ideal use-cases.

Despite Challenges, SMBs Find AI Is Paying for Itself

Snowflake’s report found that 92% of AI early adopters said their investments in the technology have already paid for themselves. Financial services organizations’ focus on financial performance helps explain their success with ROI. In addition, many organizations are leveraging the technology for low-hanging use cases improving their internal productivity, Avanes says.

Consider businesses trying to better understand their sales productivity and help their sales forces increase leads, improve outreach and messaging, and combine various data sources. They can consolidate data from Salesforce and ServiceNow via an AI-enabled platform in order to build a chatbot-like experience on top, through which sales account executives can ask questions and get answers fairly quickly. 

“If you look at what was done in the past and how would you actually put a dollar amount to it, it does translate into a pretty large win and return on investment,” Avanes says.

Part of the allure of such AI is the simplicity, particularly when an organization is fraught with legacy code that needs to be migrated, a common issues for FIs, he says.

Another attractive use case for organizations seeking immediate ROI is leveraging AI to help migrate legacy applications to a new platform, which gets complicated quickly when their data is in different formats and were written with different coding languages.

“With the newest models, that task can be well automated and really slashes the cost, the resources and time to move to a more modern platform,” Avanes says.

LEARN MORE: A new era of digital banking powered by AI technology.

Key Goals For AI Preparedness

Organizations are deploying a variety of data management approaches, but according to the research, early adopters say tools must achieve five goals:

• Break down data silos. Sixty-four percent of respondents say it’s a challenge to integrate data across sources.
• Integrate governance guardrails. Fifty-nine percent report difficulties enforcing data governance.
• Measure and monitor data quality. Fifty-nine percent of early AI adopters say measuring and monitoring data quality is difficult.
• Integrate data preparation. Fifty-eight percent say making data AI-ready is a challenge.
• Efficiently scale storage and compute. Fifty-four percent report having a hard time meeting storage capacity and computing power requirements.