Are You Avoiding GenAI as a Strategic New Technology Tool?

A recent report in The Financial Brand highlighted a report from Qualtrics that offers a 4-step approach to understanding how AI can be integrated to safely improve customer experiences. Focusing on how AI can partner with human experience to provide a robust internal or external customer experience.

The report has a helpful section where common terms related to GenAI are explained. I won’t review all of them here; you can access the report directly using the link at the end of this article. However, I do find a lot of confusion, especially among financial institution executives, about the terms used to describe GenAI. A good example is the issue of whether a GenAI tool is based on a Large Language Model (LLM) or Small Language Model (SLM). LLMs try to consume as much general information as possible. Think ChatGPT or Co-Pilot. They can get really creative in their responses, leading to inaccurate content. SLMs are trained on documents that represent a specific narrow outcome. Think of a tool that only knows your loan policies, procedures, and related loan regulations. This is very useful for those involved in underwriting and booking loans to ensure they are following the rules and policies accurately.

Here is a summary of the 4 areas that the Qualtrics paper documents:

The Five AI Customer Experience Superpowers

We tend to think of GenAI only in the Generation category. This is where we give a tool like ChatGPT a couple of sentences of instruction and it writes out a 1500-word narrative in less than 30 seconds. Every time I show someone who has never seen it work before, the only description I can give to their face is amazement. It is truly impressive. However, as previously mentioned, using that output without careful review is fraught with danger. The generated text could be misleading or plainly wrong. Its interpretation is only as good as its ability to properly understand our original intent, something that is not always provided clearly. While we may know what we want to say, that intent is not always provided clearly in the narrative we key into the dialogue box. As far as Gen AI's ability to perform optimization, I am not sure whether the popular Gen AI tools truly are coming to intelligent solutions or whether they are just trained on so many examples that they have “memorized” responses and can provide these very fast in context, making them seem prescient. More study is needed on that element.

The Human Factor: Partnering AI With People

The report highlights that there is still concern for the accuracy of AI tools. And while 27% of consumers don’t trust AI results, 45% have concerns about human interaction with AI tools, and 48% worry that AI will eliminate human-to-human contact. The report details two important segments where using AI can augment human capabilities:

I am especially enamored with the idea of using GenAI to spot anomalies in behavior. An AI tool specific to spotting abnormal events is the best, most effective way to identify suspect fraud or other similar bad outcomes. All industries, but particularly financial services, would do well to employ GenAI tools in areas where anomaly detection is mission-critical.

Four Bold AI Goals

The report highlights 4 areas where AI would have a significant impact:

If you are a frequent reader of my posts, you know that I have been on this bandwagon for many years. While generating revenue should always be a focus, the ability to use GenAI to reduce operating costs should be the initial focus. If for no other reason, the fact remains that saving a dollar is worth more than generating a dollar of revenue, since you don’t (today anyway) pay taxes on money you save.

The Human-Centric Approach to AI Implementation

The report highlights that by focusing on empowering employees and enhancing, rather than replacing, human interactions, businesses can navigate the AI landscape more effectively.

Financial institutions must stop the wholesale practice of shutting off all AI and pretending that the technology somehow doesn’t apply to FIs. GenAI can be used in specific applications today that would enhance operational efficiency and improve the customer experience. I feel strongly that areas of immediate impact are focused on internal expert systems (such as the aforementioned SLM loan database), enhanced GenAI-powered chat (don’t get me started on how bad chatbots are …), and predicting the probability of loan repayment. I am especially excited about a pilot I am running to create a GenAI agent that is focused on providing financial wisdom to younger customers/prospects. It’s not ready for prime time just yet, but I will be documenting more about this agent (who I call Buckley) in the near future. But it will all start with a reasoned approach to understanding AI and its strategic use, and you can’t get there with all AI tools shut off. An eyes-open approach must be the norm from the boardroom to the server room. Make it happen.

https://thefinancialbrand.com/news/data-analytics-banking/artificial-intelligence-banking/the-5-use-cases-for-adding-generative-ai-to-your-cx-strategy-180151/?edigestT1+2