Banks are betting on artificial intelligence to change the future of work. 

“Much of risk compliance and control depends on your understanding of what it is that people are doing, and monitoring those things,” said Avinash Misra, co-founder and CEO of Skan, a 2-year-old startup that today raised a $14 million Series A. The company offers AI-based solutions that analyze internal bank processes. “If that basic understanding is biased and wrong, then you are forever doomed to go into many cycles of either trying to transform, automate or do things that don’t work.”

As banks seek to automate back-office processes, they must contend with an amalgam of technology solutions they use daily to accomplish basic work. New AI startups, however, are taking aim at the complexity of bank functions.

San Jose, Calif.-based Skan uses computer vision, machine learning and data science to digitize the consultant role. The tech is built on the idea that, given the hodgepodge of software vendors and applications utilized by banks, employee work processes have become so nuanced that executives in charge may not even know how work gets done on the ground floor. By analyzing the time employees spend on certain screens, as well as their screen-to-screen workflow, Skan’s technology is able to identify inefficiencies. Only work-related tools set by employers are monitored and sensitive information can be redacted to avoid an invasive Big Brother feel. 

To describe the problem Skan is solving, Misra invokes Polanyi’s Paradox, the idea that humans know more than they can ever tell. “The moment something becomes a habit, or something you have already become accustomed to doing, you find it very difficult to describe it,” he said. “The act of interviewing people to find out what they are doing in a given piece of business process is doomed from the start.”

Although Misra wouldn’t name any clients, Skan is currently focused on banking and insurance companies, while also eyeing health care and pharma. With the Series A led by Cathay Innovation, Skan will hire data scientists and AI engineers to further its product development.

For big banks, companies like Skan could also serve a database function. Matt Carbonara, managing director of Citi Ventures, said most banks consist of cobbled-together ad hoc systems, creating a real need for technology like Skan. The $2.2 trillion-asset bank participated in Skan’s Series A but is not a client yet. According to Carbonara, Citi has relationships with two-thirds of the about 70 companies in its venture portfolio, and the bank would like to have a relationship with Skan down the road. 

“I see potential value in being that central nervous system of a company, where you become the system of record for their processes,” Carbonara said. “I think that’s a really valuable piece of real estate for a company like Skan.”

AI isn’t just playing a role in back-office efficiency; it’s also changing how banks interact with clients. C3.ai, an 11-year-old company that supplies big banks like Bank of America and Standard Chartered with AI solutions, announced the launch of its C3 AI customer relations management (CRM) platform this week. The Redwood City, Calif.-based company aims to help banks better predict client interactions through machine learning. 

In addition to internal bank data, C3 ingests external data like interest rates, social media and equity prices to help employees interact with retail and corporate clients. C3’s technology gives banks insights on revenue forecasting, when clients are likely to churn, and what products they should offer specific clients. The company uses machine learning, ingesting historical data on how clients fared during market fluctuations in the past, to increase the accuracy of its client predictions over time.

See also: Bank of America’s technology patents highlight AI, biometrics, augmented reality

The new C3 AI CRM runs on Microsoft Azure, taking Microsoft Dynamics 365 CRM and Adobe Experience Cloud and making them more industry-specific for fields like banking, manufacturing, telecommunications, insurance and energy. Other C3.ai clients include IBM, Shell and the U.S. Air Force. According to Ed Abbo, C3.ai’s president and chief technology officer, the company’s technology can reduce loan decision times from 70 days to a week or less, and can reduce financial crime false positives by 85%.

“A last-generation CRM system basically just says, ‘Hey you engaged this customer three months ago. Maybe you should call them again,’” Abbo said. “To really understand customer satisfaction — whether a client is going to churn or whether they are going to buy equities etc. — you really need a lot more sophistication than simply, ‘Did you speak to the client recently?’”

The 2020 Banking Automation Summit Virtual Experience, taking place November 9-10, is a new event that will provide a platform for industry professionals to share fresh insights, tools, trends and strategies. The Summit will focus on how automation will transform banking, especially with regard to back-office operations, risk management, data science and utilization, customer experience and more. Register here.