By Edwin Jongsma, VP of AI and Integrations at Financial Finesse. Financial Finesse won the ‘Best Use of AI-driven Personalization‘ category, and was a finalist in the ‘Best Use of AI for Learning’ category at The 2024 A.I. Awards.

Roughly 90% of all fintech companies are utilizing some level of AI in their operations and business models.

While most of these use cases pertain to streamlining day-to-day tasks and boosting productivity amongst employees, companies are increasingly harnessing the power of AI for the benefit of their clients, customers, and users.

Although this is an important step for the future of the fintech industry, it is essential that companies deploy AI that is safe and responsible for users, rather than integrating this technology solely for the sake of being able to say they use AI. The best way to guarantee that this technology is created safely is for the technology to be built on a closed loop model and for the human to remain in the loop.

No Room for Error

Most generative AI models are built on outside data. They learn and function by analyzing large external sources like the internet, public databases, etc. Although this can be advantageous in some scenarios, such as providing real-time updates and information for users, it can also result in an AI tool that is prone to biases and hallucinations, which can become outright dangerous, especially regarding financial guidance. Outdated and untrustworthy information might also result in inaccurate results generated by AI.

The finance industry has little room for error when it comes to the accuracy of tools and systems that users rely on for financial guidance. Missteps can result in users making decisions that harm their financial well-being and erode the trust they have in the companies deploying these tools. To safeguard against these risks, fintech companies must make it a priority to protect the financial futures of millions of users who depend on their solutions for secure and reliable coaching.

Keeping Financial Experts in the Loop

An important step in creating safe and responsible AI is to ensure the tool is grounded in the company’s own documentation and datasets. In the fintech industry, every company has a wealth of information and data at its fingertips. With an AI tool meant to impart financial guidance or suggestions, companies have the opportunity to integrate their proprietary data into the technology, ensuring that the information users receive from the tool stems from actual resources that a human financial planner or coach from the company would provide.

It is critical to build AI as a Chain of Thought system that builds in evaluations by AI and keeps human financial experts in the loop – ensuring there is a way to authenticate responses and teach the AI. By building a quality assurance program into the technology, financial coaches or planners can validate responses from the AI, helping it learn and adapt over time. Typically, this process requires heavy training on the human side upfront, but as time goes on, the tool continues to learn the difference between right and wrong responses, requiring less human interference and, ultimately, resulting in an AI tool that companies can feel confident in deploying to their clients and users.

Let’s take a look at this model in practice. Financial Finesse’s Ask Aimee Anything – an interactive financial coaching AI search engine – was built utilizing the AI Transformer technology also used by GPT but designed to overcome common AI limitations by training the resource on a closed universe of thousands of original articles and other content pieces created and maintained by our on-staff team of CFP® professional financial coaches only – not the wider internet. Additionally, financial coaches are able to interact with the quality assurance program built into the tool, monitoring the responses output by the AI and correcting any mistakes that come up. This model prevents biased or inaccurate responses, ensuring the information users receive is up to date and trustworthy.

Striking the Balance

The next challenge we face as an industry lies in finding a balance between keeping the human in the loop and letting the technology work independently. Ultimately, as the technology evolves, the goal is to reduce the time humans have to be involved but corners should never be cut.

As fintech companies move forward with integrating AI, they must recognize that responsible implementation is not just a technical challenge but an ethical imperative. The future of the industry will depend on how well companies integrate these safeguards into their technology, ensuring that innovation does not come at the expense of the user. By prioritizing accuracy and safety, the fintech industry has the opportunity to lead in demonstrating how AI can improve lives while upholding the highest standards of integrity.