AIaaS – artificial intelligence-as-a-service – is the technology discipline that eases enterprise entry into the mysteries of the AI journey while lowering the financial risk. It’s already making an impact on the financial services industry, with institutions offering AI-based retail services that keep them in ongoing contact with their customers, along with fraud detection capabilities for unmasking money laundering, among other financial sins.
At Tabor Communication’s HPC on Wall Street conference next week (Thursday, Sept. 13, Roosevelt Hotel, 45th St. and Madison Ave., New York), a data scientist and mathematician from Credit Suisse will present on the latest trends in the integration of AI and cloud computing, much of it having to do with AIaaS.
Credit Suisse’s Tassos Sarbanes shared that AIaaS in the financial services industry has evolved beyond big data analytics to real-time business use cases that are moving into the mainstream of capital markets globally. Sarbane’s keynote, scheduled for 8:15 a.m., will address the opportunities, challenges and proven approaches for developing and scaling AI for advanced business services and solutions.
According to Sarbanes, AIaaS takes three general forms, starting with serving as a personal digital assistant for customers of financial institutions. For example, Bank of America’s Erica is a popular AI-driven virtual assistant that utilizes aspects of AI, predictive analytics and natural language processing to make consumer banking transaction more efficient.
For investors, investment banks such as Goldman Sachs, Morgan Stanley and UBS Group AG of Switzerland offer “robo-advisor” platforms that advise clients on portfolio, asset and allocation management decisions. One of the most popular, Betterment, also offers certified financial planners – i.e., people – to augment AI-based advisory services for customers with accounts more than $100,000.
Sarbanes said robo-advisors are also offered by BlackRock, Vanguard, T. Rowe Price and Fidelity, with its Cora virtual reality financial agent that can interact with a client’s vocal commands.
At the enterprise level, FINRA (Financial Industry Regulatory Authority) has had FastOLA under development for several years – it’s an analytics tool for online audit trail analysis that leverages both cloud (AWS) and open source technology; it compiles large amounts of market trade data from places like NASDAQ, the National Market System, Over-the-Counter equity securities, and trades executed not on an exchange. According to FINTA, it cuts user query time by compiling data as it is received rather than compiling information on demand, reducing hours-long queries to less than a minute.
“FINRA is in full swing to catch ‘market cheaters,’” Sarbanes said. “FastOLA, running on AWS, is used internally by market regulation analysts and their management.”
Other AI-based tools used by financial institutions, such as Citi’s anti-money laundering and fraud detection tool, fight fraud by discerning whether a series of transactions could be money laundering,” Sarbanes said, “or a more innocent activity, such as a sudden wave of overseas expenses.”
For financial services companies whose developer staff or data scientists lack skills or experience necessary for adding machine learning capabilities to applications, the public cloud vendors offer fully-managed ML services, also called AutoML. “These use templates, pre-built models and/or drag-and-drop development tools to simplify and expedite the process of using a machine learning framework,” Sarbanes said.
For instance, Amazon SageMaker is a fully-managed platform that enables developers and data scientists to build, train and deploy models at various scales. The respective tool at Microsoft Azure is called Machine Learning Studio.
The major cloud vendors also offer APIs, which simplify incorporation of AI-related technologies into the application or products being built.
“Developers use the API to allow the app to access that functionality in the cloud,” said Sarbanes. “APIs are available for a wide variety of different purposes, including computer vision, computer speech, natural language processing, search, knowledge mapping, translation and motion detection.”
For example, Google Cloud offers speech, translation and vision APIs, among others.
“If implemented correctly, AlaaS can be an effective tool helping organizations augment their AI developer capabilities faster, with fewer hardware and staff resources,” said Sarbanes. “But the variety of AIaaS alternatives can be daunting. Research is required to select the right cloud service provider and to identify the right ML framework to build an app on, such TensorFlow, MxNET or PyTorch. And certain business constraints, such as security and compliance restrictions, also need to be accounted for.
“Intelligent automation will become the new normal, driving new innovations and standards. Enterprises should start finding ways to ingest implicit intelligence into their IT ecosystem. We all need to be prepared to embrace this new technology wave.”
This article originally appeared on HPCwire sister pub EnterpriseTech.