HSBC and Google Cloud team up for a regulatory-focused chatbot
March 24, 2021
HSBC chooses Google Dialogflow to improve speed and overall quality of banking policy responses.
Global banking giant HSBC and Google Cloud have teamed up again. After working together last September (2020) to improve HSBC’s Hong Kong call center service and again this February to move their blockchain custody services to the cloud, they’ve joined ranks with KPMG’s Innovation Division to design and launch Operational Resilience and Risk Application (ORRA).
ORRA, a cloud-based search-enabled chatbot, digitizes HSBC’s outdated paper trail and call-based systems to help HSBC’s global policy experts answer internal regulatory questions.
In an interview with Insider, both HSBC’s Global Head of Innovation, Finance & Risk, Steve Suarez, and Head of Risk Transformation and Innovation Lead for Asia Pacific, Gareth Butler thought the bank could use artificial intelligence (AI) to take a fresh approach to operational risk and resilience.
The aim is to give employees access to accurate policy information, enabling them to answer questions faster and with more consistency.
ORRA AND DIALOGFLOW
Built on Google’s Dialogflow, ORRA uses natural language processing (NLP) to interact with employees through natural conversations and analyzes unstructured internal document data to streamline searches for answers to local and global policy questions.
Accessible from the HSBC intranet, ORRA is available to all employees who need answers to queries about all areas of HSBC’s internal policies and frameworks. ORRA’s underlying machine learning architecture enables the chatbot to learn from user feedback to improve future responses. “As query flow increases, the architecture uses machine learning and user feedback to determine the best response to give,” explained Butler.
Dialogflow, Google’s natural language understanding platform, makes it easy to design and integrate conversational agents. It offers a rich feature set for building AI-powered chatbots and natural-language conversational agents that can scale to support millions of users.
Using Dialogflow’s flexible and easy-to-use coding model, ORRA’s development team was able to build a scalable, conversational platform with document search capability in just four months. Native features like small talk enabled them to create an agent that provides a natural conversational experience to users, while training tools help the chatbot learn from customer feedback to provide better responses.
Dialogflow, built on secure cloud infrastructure, presents a fast and reliable way for financial institutions to stay connected to customers while ensuring security and compliance.
“HSBC is a highly regulated organization, so managing risk effectively and efficiently is critical to making sure that we keep our customers safe and give them a great customer experience,”
CONCERNS ABOUT SECURITY SLOW ADOPTION
Typically, financial institutions have been slow to adopt the public cloud because of concerns about security and compliance. But advances in technology and the lure of potential cost savings have seen more and more institutions become comfortable enough to make the leap.
The COVID-19 pandemic has also played a part in accelerating acceptance. In early 2020, PayPal migrated key parts of its payment infrastructure to the cloud, and more recently, Capital One closed its last data center to move entirely to a cloud environment. With customers unable to do business physically, conversational chatbots present a helpful and friendly alternative to face-to-face interaction.
But so far, except for JPMorgan Chase’s implementation of COIN to streamline back-office operations, customer-facing scenarios have been the more highly adopted use cases for AI chatbots in the financial sector. These include enhancing customer experience, lead generation, customer onboarding, and support for financial services like budgeting tips and loan applications.
“In the future, we are going to be interacting with chatbots frequently for routine transactions. As a leader in the financial services industry and a technology innovator, HSBC is taking the first steps in using cloud-based chatbot technology to get fast, accurate answers to our customers.”
HSBC’s implementation of ORRA to handle regulatory compliance controls represents a completely new use case in the financial sector. Using cloud technologies like AI and machine learning not only helps employees to respond rapidly and accurately to policy-related queries but also gives financial institutions the flexibility to adapt to complex and ever-changing regulatory requirements.
According to Chris Wilson, HSBC’s head of Architecture, Policy & Regulatory Mapping, future versions of ORRA will include guidance on judgment calls and more functionality related to risk acceptance, risk issues, and risk relevance.
As ORRA handles routine queries from internal employees, subject-matter experts can focus on more complex policy issues. Big data analysis can also help financial institutions identify market trends, streamline procedures and reduce risk.
With technologies like ORRA, financial institutions can not only create conversational agents that take the pressure off of their internal resources but gather information through analytics and learning that can also inform future business decisions and policy and procedural changes.
FUTURE AUTOMATION IN THE FINANCIAL INDUSTRY
Trends indicate that while human advisors continue to play a vital role, automation through AI development is transforming the financial industry. A recent PricewaterhouseCooper study shows that 52% of financial services decision-makers are making substantial investments in artificial intelligence (AI).
As more information is stored in the cloud, the financial sector can deliver more advanced interactions at a lesser cost. And while front office implementations that deepen customer relations continue to offer the greatest cost-benefit, middle and back-office AI applications present an opportunity to use automation for risk assessment, fraud management, and to improve anti-money laundering (AML) processes.
“As we move forward, we’ll be adding in more around risk acceptance, risk issue, and risk relevance.”
The future of AI in banking not only involves personalized banking experiences for customers but automated agents perform middle and back-office activities, assist with regulatory and compliance policies and procedures while gathering information that informs business decisions.
As more advanced use cases emerge, automated AI applications will evolve to help financial institutions adapt and innovate to unlock new paradigms.
Butler told Insider that future versions of ORRA will accommodate more policies and documents and support multiple languages, as well as mobile interactions and speech. As he rightly pointed out, “This is only the beginning of our conversational AI journey”.