Organizations like Home Depot and Lowes grew exponentially over the past decades due to an emphasis on DIY or “Do It Yourself”. Service at those big box stores is a challenge (how you are supposed to identify and find the correct nozzle for a sprinkler head), but there is no question about the interest we Americans have in the DIY concept. Even online shopping services like Amazon allow DIY to buy pieces and parts needed for a repair. eBay also has a robust auto parts business, again perfect for the home mechanic. Yet, how has the DIY concept translated over to financial services? The availability of robust digital banking services has enabled users to manage virtually every aspect of their financial relationship online. This is truly the essence of DIY. I can deposit checks, transfer funds, pay bills, access faster payments and much more, all from my computer or mobile device. However, there is a growing trend in some financial services known as DIFM or “Do It For Me”. Now you might be thinking this is merely returning to the place where a financial institution would perform transactions on behalf of a customer request at a drive through, call center or branch location. But DIFM is distinctly different in that it is an autonomous system acting automatically on behalf of a customer without any human intervention. Let’s examine how it can be used. For my regular checking account, I have an alert set up that will trigger a notification if my account balance goes below $500. If I receive that alert, I can make a decision about whether I want to transfer funds and if so, which internal or external account I would use. This is a DIY model. I setup the alert, and when the alert is triggered, I am making a decision to transfer or not. There is no AI or Machine Learning (ML) involved. Now, consider if there was a level of automation available from my financial institution that was intelligent enough to make an autonomous decision. Powered by AI and ML, this tool would evaluate many factors including guidance I might have provided upon activation, reviewing past transfer history that aligns with a below balance event and other factors such as knowledge of a pending or upcoming recurring credit. This tool would see the account balance go below $500 and based on its previous learning, automatically make a decision to transfer funds to ensure there is no overdraft. This is DIFM. My question is how would you feel about offering your customer such a service? The technological ability to perform the automated scenario I describe above is fast approaching. Dubbed Agentic AI, it combines Generative AI, Machine Learning and enterprise level automation to achieve the ability to determine and execute actions autonomously. In a true Agenic AI mode, I would not receive advance notice of the actions Agenic AI Agent is executing on my behalf. I would still have access to my account information at my discretion and could manually execute a transaction. I might even get a text or email with acknowledgement of the action performed by the Agent, but neither me nor anyone at my institution would be in the chain of creating or authorizing the action. It is autonomous. There is a growing set of current and future customers who likely desire this type of service, perhaps representing younger customers who have DIFM interactions with other digital services. But bankers might think that this brings up potential problems, such as issues of liability, where the bank’s autonomous service is performing consumer transactions. Is there an opportunity for the Agent to act in a manner that is not consistent with the customer’s intent? How will you explain to an auditor or examiner what is going on in the “black box” of the Agenic AI Tool? And does the Agenic AI Agent concept really represent an improved service over the self-serve nature of the digital banking suite we currently offer. Don’t customers, even young millennials and Gen Zs, think about privacy and control over their money differently than entertainment, gaming and chat? Nearly every study I’ve seen on younger customers indicates their interest in learning good money management skills and looking to their financial institution to provide education and financial wisdom. Based on this finding, it would seem they would be initially accepting of a DIFM Agent. However, once actively using it, they realize that each scenario is somewhat unique. After all, an Agent is not going to know that your buddy is repaying you $200 this afternoon that you are depositing to your account. Like most things, there is the promise of what an Agenic AI Agent can do and the reality of how that gets implemented into a consumer facing digital service. Perhaps there is some middle ground we could examine. In a previous post, I highlighted my exploration of an interactive concierge tool I named Buckley (you can read that blog post here ). I proposed that Buckley could always be “on” evaluating your accounts, transactions and balances. Using my earlier example, Buckley would advise that the account balance had dipped below the $500 level, evaluate all of the relevant data and make a specific transaction recommendation. I could then verbally accept Buckley’s recommendation or provide an alternative action or tell Buckley to ignore the situation. This would give me the full benefit of the robust analytics from an Agenic AI Agent without the autonomous action. I think the long-term benefits of this model would be best appreciated by customers without the risks inherent in a full DIFM model based on an Agenic AI Agent. Perhaps we’ll call it “You Make the Call Based on My Wisdom” or YMTCBOMW. Hmmm, that’s a big mouthful … I got to work on a better moniker. If you have an idea of what to call my idea, contact me at dpeterson@bankers-bank.com. Thanks in advance for any marketing gems you can pass my way … Resources https://thefinancialbrand.com/news/artificial-intelligence-banking/agentic-ai-the-next-big-innovation-for-banks-186428 The views expressed in this blog are for informational purposes. All information shared should be independently evaluated as to its applicability or efficacy. FNBB does not endorse, recommend or promote any specific service or company that may be named or implied in any blog post.