by Debashis Rana –
Financial Services firms have pioneered the use of technology and data since the 1980s. A few use cases that provide a glimpse of the ‘state of the industry’ vis-à-vis AI and related disciplines include:
- Chat-bots help customers with routine tasks and warn of unusual spending
- Deep learning ranks stocks to better inform investment strategies
- Machine learning helps identify unusual user behavior inside and outside the firewall to detect fraud
- Semantic technology provides context-based inferences for know your customer (KYC) and anti-money laundering (AML) efforts by banks
- Location intelligence enhances the understanding of risk by mortgage lenders
- Non-linear relationships improve risk mitigation for asset managers
- High-frequency trading algorithms continue to improve
It is important to note that the above represent a small sample. There are too many to list exhaustively, even assuming that were possible.
Significant trends in the use of AI in Financial Services include:
- While supervised learning is still the order of the day for most AI solutions, reinforcement learning and unsupervised learning are making small but noticeable inroads in use cases that are not mission-critical or ‘life-or-death.’
- Alternative data continues to enjoy a steady increase in adoption; among these social media data is the heavy hitter as firms step up their digital engagement with younger audiences as part of their digital transformation efforts.
- Cyber-security capabilities have gotten a significant boost. However, automation has been a double-edged sword with bad actors launching increasingly sophisticated attacks that need constant innovation to stay ahead.
- Regulatory uncertainty is causing some delays in adoption due to perceived risk, however, surveys show that as regulations crystallize in the long term, they will reduce overall risk across the industry.
- Data-set bias propagated by models is an area of concern, especially in credit analytics; early predictions of reduced bias of machine-based reasoning have not borne fruit, but reinforcement learning is seen as a primary solution while explainable AI catches up.
Putting AI at the Financial Services firm’s core
Top priorities for Financial Services firms today are Customer Engagement and Operational Efficiency. Let’s examine what this means for the application of AI.
As Financial Services firms continue to pivot from product- to customer-centricity, the biggest differentiators revolve around how the customer is treated throughout the lifetime of their relationship with you. This covers the gamut from attracting, acquiring, serving to anticipating their needs and wants. Each of these relationship stages has unique opportunities that are greatly benefited by AI-based thinking and, therefore, solutions. For example:
- Attract: use alternative data sources to know more about a prospect, especially in early digital interactions – what can you tell them, predict for them, and offer while you still ‘officially’ know very little (e.g., they have visited your web site a handful of times)
- Acquire: use historical buying behavior patterns of similar prospects/customers in conjunction with what you’ve learned in the ‘attract’ phase to convince them to buy (or buy more, a.k.a. up-sell)
- Serve: critically examine all customer-facing business processes to automate rote (which are usually well-defined) steps, examine them again in conjunction with customer behavior and feedback continuously to further accelerate your business processes
- Anticipate: don’t wait for the customer to tell you what to do, use the power of AI to predict what they need or want next and act on it; if you make a wrong prediction, treat it as you would any other negative customer behavior or feedback, and learn from it
Staying competitive by increasing the top line and reducing the bottom line has been a reality since the start of commerce. Financial Services firms these days are under tremendous pressure from fintech. It is especially tricky since fintech has pioneered AI in the industry, and traditional firms are playing catch up. The areas of opportunity that AI can enable or improve are broad, and the salient ones are as follows:
- Virtual Assistants: reduction in interactions that require a human in the contact center need to evolve further to span all functions – marketing, selling, underwriting, servicing, complaints, compliance and more
- Relationship Management: as firms diversify to stay relevant in the marketplace, arming employees with contextually relevant information in real-time is key so that they can focus on the customer instead of expertise in a broadening product portfolio
- Fraud Detection: traditional rule-based methods are being augmented by pattern recognition and machine learning, to the point where more sophisticated types of fraud are being detected, thereby reducing financial and reputation loss
- Regulatory Compliance: In a landscape where lookalikes of GDPR have been adopted in the United States (CCPA) or will be imminently adopted, contemporary technology is being used to its full advantage to ensure and report compliance
Building AIQ and Competence
So, how do you set your firm on this path to success with AI? The answers sometimes come from the outside rather than inside the firm. This doesn’t imply that the experience of your staff isn’t valuable. It’s just that the thinking required to build AIQ and stay ahead of the competition using it are not necessarily native skills. Early adopters and mature players alike have leveraged some or all of the following:
- Incubators – Bring external thought leaders and selected staff (typically change agents, forward thinkers, and high-energy employees) together in an innovation-oriented offsite setting; bring real problems and create real AI-enabled solutions that can be prototyped during the session.
The goal should be to take back something of value to the firm that can be operationalized in a short time-frame.
- Ambassadors – Recruit and incentivize willing influencers who can push the envelope to provide and brainstorm real-life ‘edge cases’ that require going the extra mile (think AI) beyond traditional solutions. Besides, laying bare some of your front-, mid- and back-office to this select audience will make them appreciate what you do for them and spread the word.
The goal should be to take back something that you hadn’t thought of for your next iteration, or to your next Incubator.
- Grooming – Use AI to study the ongoing work behaviors of your change agents, forward thinkers and high-energy employees; in other words, ‘drink your own champagne’ to see if their thinking is indeed improving as a result of participating in Incubators. Studies have shown that the best AI Designers blend user experience with engineering and non-traditional thinking to achieve the ‘art of the possible.’
The goal should be to arrive at a set of skills that you need to hone in house and look for in new hires.
- Outsourcing – Use professional services expertise to set yourself down the path, or elevate to the next plateau, of AIQ and competence; the benefits of this approach are that you will be piggybacking on others’ mistakes and could save valuable resources, time and money.
The goal should be to achieve one of the following:
- AI as a Service – a hands-off outcomes-driven capability with a trusted partner
- Co-development – a risk-sharing agreement with a trusted partner
- Insourcing – an organized transition from AIaaS to co-development to in house
Treat AI as the latest addition to your arsenal as you continue to innovate to deliver value to your customers and your organization. The critical takeaways to applying AI are:
- Strategic alignment of AI with business goals is a prerequisite – without it, all efforts are ‘science projects’ that will provide little to no value
- Outcomes should drive priorities – insights alone are not sufficient; they need to tie to actions towards business goals with measurable ROI
- Automation provides plenty of opportunities for acceleration – use it to first identify and then mechanize relevant parts of your business processes
- Feedback loops are critical to continued success – end-to-end measurements of AI-enabled processes enable ROI and continuous improvement
- Alternative data provide invaluable raw material for new methods – use AI with companion disciplines like location intelligence to augment traditional techniques
- Think outside the box when it comes to things that are important to the business – AI provides plenty of food for thought in this regard
Financial Services firms are an early adopter of everything; AI is no exception. While the numbers vary, several studies agree on one thing: at least 50% of Financial Services firms have AI as part of their portfolio for 2020. The time to adopt (if you haven’t done so already) or improve your capability is now.