The Buzz – Details, episodes & analysis

Financial institutions are moving beyond pilot projects to implement production-grade, explainable and cost-effective AI solutions that can meet operational and regulatory demands. 

AI has evolved rapidly since fintech Arteria AI was founded in 2020, Amir Hajian, chief science officer, tells Bank Automation News in this episode of “The Buzz” podcast. The company provides banks with AI-powered digital documentation services. 

“2020 was a very simple year where AI was classification and extraction, and now we have all the glory of AI systems that can do things for you and with you,” Hajian says.  

“We realized one day in 2021 that using language alone is not enough to solve [today’s] problems.” The company began using multimodal models that can not only read but search for visual cues in documents. 

AI budgets and strategies vary widely among FIs, Hajian says. Therefore, Arteria’s approach involves reengineering large AI models to be smaller and more cost-effective, able to run in any environment without requiring massive computer resources. This allows smaller institutions to access advanced AI without extensive infrastructure. 

Hajian, who joined Arteria AI in 2020, is also head of the fintech’s research arm, Arteria Cafe. 

One of Arteria Cafe’s first developments since its creation in January is GraphiT — a tool for encoding graphs into text and optimizing large language model prompts for graph prediction tasks.  

GraphiT enables graph-based analysis with minimal training data, ideal for compliance and financial services where data is limited and regulations shift quickly. The GraphiT solution operates at roughly one-tenth the cost of previously known methods, Hajian says. 

Key uses include: 

• Regulatory compliance; 
• Anti-money laundering; 
• Predictive modeling; and 
• Low-data problem solving. 

Arteria plans to roll out GraphiT at the ACM Web Conference 2025 in Sydney this month. 

Citizens Bank is advancing its digital payments strategy in 2025, with a focus on embedded finance and omnichannel capabilities. 

For example, Citizens is moving from screen scraping to an API-enabled environment, allowing it to use insights from its private banking segment in other segments, such as SMB, mid-market and commercial, Taira Hall, executive vice president and head of enterprise payments at the $217.5 billion Citizens, tells Bank Automation News in this episode of “The Buzz” podcast. 

“One of the benefits of having an enterprise payments function is … we can look at all of our core payment rails and understand at scale … It’s sort of bringing this intertwined aspect to it … making sure that anything we build in one place has connectivity tissue into the rest of the bank,” she says. 

The bank, she says, is moving forward with its multiyear digital transformation, focusing on: 

• Fintech partnerships; 
• Data and analytics investment; 
• AI deployment; and  
• Expanded capabilities for existing tools, such as the bank’s SMB-focused Cash Flow Essentials payments platform. 

Learn more from Hall about how Citizens is looking to balance relationship banking with digital innovation in this episode of “The Buzz.”

Subscribe to The Buzz Podcast on iTunes or Spotify, or download the episode.

Listen as Matt Denham discusses how fintechs can boost digital engagement on FI platforms.  

After a financial institution has successfully digitized its loan origination and account opening system, improving the user experience is the next step. Listen as Kristiane Mandraki, director of business development and marketing at software developer fintech Praxent, details how financial institutions can leverage automation in their back-end processes to enhance their customer experience.

Businesses looking for quick credit-decisioning can benefit from financial reporting that gives banks a clear picture of data. Listen as Sid Saxena, chief executive of artificial intelligence (AI)-powered accounting automation platform Docyt discusses the technology, its integration and how lenders can benefit from automated accounting to streamline decision-making.

The $1 billion Cogent Bank recently integrated blockchain-based, real-time payment platform TassatPay to provide its business-to-business (B2B) clients with 24/7 instant payment options.

Cybersecurity measures are a priority at Royal Bank of Canada (RBC), from monitoring and mitigation to solution investment.

Banks rarely consider cybersecurity a finished process; the nimbleness of fraudsters and quick development of new hacking technology spur perpetually evolving security measures for risk and anti-money laundering (AML) divisions at most large financial institutions. But keeping cybersecurity agile goes far beyond yearly tweaks and minor investments, Shekher Puri, vice president of digital components and platforms at $1.4 trillion RBC, tells Bank Automation News in this episode of “The Buzz” podcast.

Cybersecurity is “an investment in perpetuity,” Puri says. “This is millions of dollars every single year that we're putting in and investing into these various technologies. Some of them are proprietary, and some of them we're going out to the marketplace, and we're looking at the best vendor solutions and the best supplier solutions.”

Adding nuance to this “full-court press” approach, Puri tells BAN that tackling cybersecurity at the solution architecture stage of development is key to a top-quality technology and strategy stack.

“Make sure that cybersecurity is built up a lot higher in the process,” Puri says. “They should be at the table as you're designing your product, as you're designing your experience, as you're designing your flow. They should not be an afterthought.”

Listen as Puri details RBC’s cybersecurity and risk management approach for BAN, and discusses best practices for smaller banks embarking on the cybersecurity journey.

Synthetic identities are increasingly used to attack buy now, pay later (BNPL) offerings, says Mike Cook, vice president of fraud solutions commercialization at identity verification firm Socure.  

“Synthetic identities take a different bunch of different forms, but basically, they're just identities that don't really exist,” Cook tells Bank Automation News in this episode of “The Buzz” podcast. “Fraudsters will always attack a new vertical.” 

Socure works with five of the six biggest banks in the U.S., including Wells Fargo and Capital One, as well as digital banks Chime and SoFi.  

BNPL is an attractive target for cybercriminals since the loans are submitted with little information for quick approval, making it challenging for companies to filter out fraudulent applications, Cook says on the podcast. 

There are two main ways to launch a synthetic fraud attack, he says, including “credit washing” one’s own credit report to artificially inflate the person’s credit rating and fabricating an account to open credit lines the fraudster has no intention of paying back.

Customer service systems often automate call recording and processing analysis, but you can tell a lot by detecting just the tone of a conversation. 

 

Los Angeles-based Behavioral Signals says its technology is more effective than humans at determining empathy and intention. Rather than attempting to detect the actual words being said and how they’re used, there is the “acoustic” content of a conversation, Chief Executive Rana Gujral tells Bank Automation News in this episode of “The Buzz” podcast, which provides a brief intro to the technology. 

 

Behavioral Signals was founded in 2016 and with its tone determination technology, the company seeks to match each customer with the right service employee. 

 

“In a conversation, there are two big elements,” Gujral says. “One is the spoken word, and then everything else that’s behind the spoken word or beyond the spoken word.” 

 

Behavioral Signals’ technology measures conversational elements in real time, he tells BAN, and can place information in “buckets” of emotions, behaviors and propensities. From those elements and indicators, the company’s systems can determine tone and empathy 15% more accurately than a human. 

 

Bank Automation Summit Fall 2022, taking place Sept. 19-20 in Seattle, is a crucial event on automation and automation technology in banking. Learn more and register for Bank Automation Summit Fall 2022. 

 

Use cases for intelligent document processing (IDP) include automating the know your customer (KYC) process. 

The KYC process previously was performed by humans, which led to plenty of effort and errors, says Sateesh Seetharamiah, director at artificial intelligence and automation company EdgeVerve. Using IDP, one “very large U.S. financial bank” is now able to process 20 million documents through automation, he tells Bank Automation News, without naming the bank. 

The bank “during COVID times faced tremendous challenge, and now this has become the way of the business where we are helping them in the KYC process,” Seetharamiah told BAN. “There was a lot of human effort on this and historically, they felt there were errors, and all of that has been eliminated significantly.” 

IDP leverages a multitude of technologies to automate document processing, including machine learning, natural language processing and, on the back end, robotic process automation for integration.  

Here, Seetharamiah shares how the technology can automate and add value for financial institutions with: 

• Customer onboarding; 
• Loan processing; 
• Loan dispersal; 
• Customer profile changes;  
• Regulatory compliance, including KYC; and 
• Risk mitigation. 

Listen as Seetharamiah also offers advice for how financial institutions can get started and successfully implement IDP solutions.