In 1928, many women who were looking to slim down turned to the Battle Creek Health Builder — a vibrating belt machine placed around an individual’s waist that purported to jiggle away fat without its user having to exert any effort. Though vibrating belt machines were eventually proven to be completely useless, they remained popular until the late 1960s and are just one of many passive exercise machines that have been embraced by the public over the years. Nearly 100 years later, there are multiple passive exercise gadgets that continue to be marketed to the public.While automated exercise equipment has never lived up to the hype, financial institutions — as well as many other organizations — are realizing that agentic artificial intelligence (agentic AI) is more than just smoke and mirrors. By now, there are few people unfamiliar with AI — software that enables machines to mimic human behavior and decision making by analyzing massive quantities of data. In the few years since generative AI has become popular, there have been enough advancements that the technology already has the capacity to foster its own ongoing advancement and sophistication. Enter, agentic AI.
What is agentic AI?Unlike generative AI, which relies on data-based algorithms to generate new content such as images, text and videos, agentic AI mimics human decision making by making choices, taking actions, and adapting to changing circumstances or information to complete tasks without human assistance. Using a combination of technologies — natural language processing, machine learning, and automation — agentic AI can autonomously execute elaborate activity sequences, scour databases, and complete activities. Agentic AI is also developing the ability to discern which online sources are most reliable, reducing the likelihood of inaccurate outcomes due to faulty data sometimes experienced by generative AI, which are also known as “hallucinations.” This means that organizations can use agentic AI for specialized tasks, an idea that is garnering significant interest as many businesses struggle with shortages of specialized labor and specific skill sets. Agentic AI roles can also be created quickly, providing organizations the ability to design multiple specialties as needed. For example, agentic AI could be used to assist with customer service. Agentic AI can benefit financial institutions.Instead of basic chatbots equipped to provide simple responses to the most commonly asked questions, agentic AI can be used to scale the abilities of chatbots to a whole new level. Here are some possible use cases for customer-facing capabilities:
What is agentic AI?Unlike generative AI, which relies on data-based algorithms to generate new content such as images, text and videos, agentic AI mimics human decision making by making choices, taking actions, and adapting to changing circumstances or information to complete tasks without human assistance. Using a combination of technologies — natural language processing, machine learning, and automation — agentic AI can autonomously execute elaborate activity sequences, scour databases, and complete activities. Agentic AI is also developing the ability to discern which online sources are most reliable, reducing the likelihood of inaccurate outcomes due to faulty data sometimes experienced by generative AI, which are also known as “hallucinations.” This means that organizations can use agentic AI for specialized tasks, an idea that is garnering significant interest as many businesses struggle with shortages of specialized labor and specific skill sets. Agentic AI roles can also be created quickly, providing organizations the ability to design multiple specialties as needed. For example, agentic AI could be used to assist with customer service. Agentic AI can benefit financial institutions.Instead of basic chatbots equipped to provide simple responses to the most commonly asked questions, agentic AI can be used to scale the abilities of chatbots to a whole new level. Here are some possible use cases for customer-facing capabilities:
- Providing personalized answers based on real-time knowledge of an individual’s complete financials, including total outstanding debts and available financing options to pay them off.
- Initiating financial transactions when prompted by a customer, while intelligently validating details and potentially identifying cost-saving opportunities.
Agentic AI could benefit financial institutions internally as well. Here are a few examples:
- Monitoring risks. Since agentic AI can analyze massive quantities of real-time data, it can be used to identify trends or market changes that signal the need to adjust investment strategies or holdings.
- Supplementing compliance efforts. On the compliance front, agentic AI could be used to help monitor overall risk advisory models, as well as to supplement advisory models with information and components unique to individual customer behaviors.
Here’s how other industries are using agentic AI.Outside of the banking industry, uses for agentic AI extend to things such as supply chain management, where predictive assessments can flag when certain products are likely to run out based on inventory levels. In the case of individual orders, when deliveries are going to be late, agentic AI can preemptively inform customers and offer additional discounts to keep them happy. Many organizations have already begun using the technology internally as well, often to streamline processes that were previously extremely time consuming. Here are a few ways agentic AI is being used by tech companies:
- HR resources. Cybersecurity company Palo Alto Networks has adopted an internal agentic AI program to provide human resource support for employees. The system, which interacts conversationally with employees, provides almost instantaneous personalized responses to employee questions and issues.
- IT support. Software company Jamf uses agentic AI to supplement IT support for its employees. Instead of having to contact IT support people for software needs, Jamf’s employees now rely on an AI copilot called Caspernicus, which has become the go-to for more than 70% of the company’s workforce.
Beware of the risks of agentic AI.There is no doubt that community financial institutions should familiarize themselves with the ins and outs of agentic AI but, despite its potential benefits, both internally and on the customer-facing side, diving in too quickly is not a good idea. As with generative AI, there are widespread concerns about the impact that adopting agentic AI will have on labor forces. Privacy is another major concern, given the copious amounts of data the applications rely on, especially when that includes individuals’ personal financial information. Likewise, agentic AI usage could also increase cybersecurity and regulatory risks, as regulators have yet to determine how oversight and regulation of AI will play out — particularly regarding financial institutions. Finally, since agentic AI models may rely on generative AI sources, there is always the risk that some of the data being incorporated into models may not be accurate. While community financial institutions may be eager to learn how to leverage agentic AI, it is still too early to know how this new technology will be adopted by the banking industry and whether using it will open organizations to new regulatory or cybersecurity risks. There are numerous potential benefits of embracing agentic AI, but there is still a lack of clarity regarding the potential drawbacks.
