15 questions FI leaders ask about agentic AI
Citi, Nvidia, Temenos weigh in
Manual processes still play a role in the financial services industry, making them ripe for change.
And though Mastercard, Citi and Temenos are among the financial institutions that have deployed agentic AI, the technology will become more mainstream as it becomes more trusted and as returns on investment are realized.
“Most agentic AI projects right now are early-stage experiments or proof of concepts that are mostly driven by hype and are often misapplied,” Anushree Verma, senior director analyst at think tank Gartner, said in a June 25 release. Companies “need to cut through the hype to make careful, strategic decisions about where and how they apply this emerging technology.”
Agentic AI holds great promise for FIs to gain efficiency, but building trust and regulatory clarity are among the top hurdles that must be overcome, Bentzi Aviv, global head of customer fintech solution provider at Amdocs, told Bank Automation News.
“Unlike older tech like [robotic process automations], agentic AI can make autonomous decisions to reduce human involvement, which raises concerns over mistakes, but regular accuracy and well-functioning models will overcome such issues,” he said.
The industry still uses many manual processes, Jeremy Ung, chief technology officer at software provider Blackline, said.
At FIs, there is still “time spent doing tasks that are repetitive and that are ripe for automation,” he said.
While agentic AI continues to draw attention, leaders must consider opportunity and risk before diving headfirst into the tech.
READ MORE: The next frontier in financial services: Agentic AI
BAN asked business technology leaders what questions they ask before deploying agentic AI. Leaders from $1.6 trillion Citigroup, chipmaker Nvidia, payments service provider Boost Payments, data giant Moody’s Analytics, $61 billion Valley Bank, banking technology provider Temenos, cloud service provider Amazon Web Services, fintech Amdocs and software provider Blackline weighed in on their considerations.
Here are 15 questions tech leaders are asking:
- What are the top use cases suited for agentic AI acceleration?
- How does the company maintain human touch in loop oversight with AI agents?
- Is there a sufficiently detailed understanding of the process that is being considered for improvement with agentic AI?
- Is the content that will train the agentic AI solutions of sufficient quality, governed appropriately and in the right format?
- Is the organization ready for the changes needed to derive value from agentic AI?
- How will agentic solutions be maintained, scaled and integrated within the broader technological infrastructure?
- How will the company ensure transparency of the whole process?
- What data and integrations are needed to support agentic workflows?
- Is the data required for agentic AI complete, verifiable and in a standardized and consistent format?
- Is the right talent in place to support and scale agentic AI initiatives?
- Have the systems and workflows been thoroughly assessed to identify where agentic AI can deliver the most value?
- What is the expected return on investment, and how will success be measured?
- Will this initiative distract from or strengthen the core business?
- How will agentic AI adoption position the organization competitively?
- How will audit requirements be satisfied if agentic AI is implemented?
Register here for “Seamless integration: The new frontier in embedded payments,” a free Bank Automation News webinar set for Tuesday, July 22, at 11 a.m. ET.
