I don’t know about you, but when I think of artificial intelligence, I think Steven Spielberg and Arnold. That was until I saw a solution offered by Conversica, a Salesforce partner.
AI is here, it’s happening now and it’s a lot more pervasive than you think. The rise of “robo advisers” in financial services, Ikea’s “Anna” customer service rep and Alaska Airline’s “Jenn” all point to the growing adoption of technology that personalizes customer experiences….at scale.
One of the 5 D’s of Disruption in insurance is “Dialogue.” And AI is driving it.
Today, in insurance, AI is used to create natural dialogue with customers, nurture those leads, prioritize them for agents and follow through as needed. Conversica, for example, gets smarter as it interacts more with customers. And, yes, it has passed the Turing test.
It is particularly well-suited for B2C because the volume of interactions with prospects can be overwhelming for insurance agents. As insurers embrace omni-channel, new prospects can be created from any source, whether it be a contact center, social media or a face-to-face meeting. Not only is lead volume increasing, but it takes as many as six before an agent can get a prospect on the phone. This becomes a time and energy suck for agents; he is unable to follow through on every lead, and the quality of interactions goes down.
So how are insurers and agents responding? In this webinar, Eric (Conversica) and Alex (Spring Venture Group) explain to me how AI is used to nurture and convert leads.
My takeaway: AI is not just a science project. It works. It’ll become more invisible to consumers. And it creates real value to both customers and employees.
As Marc Benioff, CEO of Salesforce, said recently in Fortune magazine, “We’re in an AI spring. I think for every company, the revolution in data science will fundamentally change how we run our business because we’re going to have computers aiding us in how we’re interacting with our customers.”
Long-time Costa Rican National Champion Bernal Gonzalez told a very young me in 1994 that the world’s best chess-playing computer wasn’t quite strong enough to be among the top 100 players in the world.
Technology can advance exponentially, and just three years later world champion Garry Kasparov was defeated by IBM’s chess playing supercomputer Deep Blue. But chess is a game of logic where all potential moves are sharply defined and a powerful enough computer can simulate many moves ahead.
Things got much more interesting in 2011, when IBM’s Jeopardy-playing computer Watson defeated Ken Jennings, who held the record of winning 74 Jeopardy matches in a row, and Brad Rutter, who has won the most money on the show. Winning at Jeopardy required Watson to understand clues in natural spoken language, learn from its own mistakes, buzz in and answer in natural language faster than the best Jeopardy-playing humans. According to IBM, ”more than 100 different techniques are used to analyze natural language, identify sources, find and generate hypotheses, find and score evidence and merge and rank hypotheses.” Now that’s impressive — and much more worrisome for those employed as knowledge workers.
What do game-playing computers have to do with white collar, knowledge jobs? Well, Big Blue didn’t spend $1 billion developing Watson just to win a million bucks playing Jeopardy. It was a proof of concept and a marketing move. A computer that can understand and respond in natural language can be adapted to do things we currently use white collar, educated workers to do, starting with automating call centers and, sooner rather than later, moving on up to more complex, higher-level roles, just like we have seen with automation of blue collar jobs.
In the four years since its Jeopardy success, Watson has continued advancing and is now being used for legal research and to help hospitals provide better care. And Watson is just getting started. Up until very recently, the cost of using this type of technology was in the millions of dollars, making it unlikely that any but the largest companies could make the business case to replace knowledge jobs with AIs (artificial intelligence). In late 2013, IBM put Watson “on the cloud,” meaning that you can now rent Watson time without having to buy the very expensive servers.
Watson is cool but requires up-front programming of apps for very specific activities and, while incredibly smart, lacks any sort of emotional intelligence, making it uncomfortable for regular people to deal with it. In other words, even if you spent the millions of dollars to automate your call center with Watson, it wouldn’t be able to connect with your customer, because it has no sense of emotions. It would be like having Data answering your phones.
Then came Amelia…
Amelia is an AI platform that aims to automate business processes that up until now had required educated human labor. She’s different from Watson in many ways that make her much better-suited to actually replace you at the office. According to IPsoft, Amelia aims at working alongside humans to “shoulder the burden of tedious, often laborious tasks.”
She doesn’t require expensive up-front programming to learn how to do a task and is hosted on the cloud, so there is no need to buy million-dollar servers. To train her, you literally feed her your entire set of employee training manuals, and she reads and digests them in a matter of a few seconds. Literally, just upload the text files, and she can grasp the implications and apply logic to make connections between the concepts. Once she has that, she can start working customer emails and phone calls and even recognize what she doesn’t know and search the Internet and the company’s intranet to find an answer. If she can’t find an answer, then she’ll transfer the customer to a human employee for help. You can even let her listen to any conversations she doesn’t handle herself, and she literally learns how to do the job from the existing staff, like a new employee would, except exponentially faster and with perfect memory. She also is fluent in 20 languages.
Like Watson, Amelia learns from every interaction and builds a mind-map that eventually is able to handle just about anything your staff handled before. Her most significant advantage is that Amelia has an emotional component to go with her super brains. She draws on research in the field of affective computing, “the study of the interaction between humans and computing systems capable of detecting and responding to the user’s emotional state.” Amelia can read your facial expressions, gestures, speech and even the rhythm of your keystrokes to understand your emotional state, and she can respond accordingly in a way that will make you feel better. Her EQ is modeled in a three-dimensional space of pleasure, arousal and dominance through a modeling system called PAD. If you’re starting to think this is mind-blowing, you are correct!
The magic is in the context. Instead of deciphering words like insurance jargon when a policyholder calls in to add a vehicle or change an address, IPsoft explains that Amelia will engage with the actual question asked. For example, Amelia would understand the same requests that are phrased different but essentially mean the same thing: “My address changed” and “I need to change my address.” Or, “I want to increase my BI limits” and “I need to increase my bodily injury limits”.
Amelia was unveiled in late 2014, after a secretive 16-year-long development process, and is now being tested in the real world at companies like Shell Oil, Accenture, NNT Group and Baker Hughes on a variety of tasks from overseeing a help desk to advising remote workers in the field.
“A large part of your brain is shackled by the boredom and drudgery of everyday existence. […] But imagine if technology could come along and take care of all the mundane chores for you, and allow you to indulge in the forms of creative expression that only the human brain can indulge in. What a beautiful world we would be able to create around us.”
His vision sounds noble, but the reality is that most of the employees whose jobs get automated away by Watson, Amelia and their successors, won’t be able to make the move to higher-level, less mundane and less routine tasks. If you think about it, a big percentage of white collar workers have largely repetitive service type jobs. And even those of us in higher-level roles will eventually get automated out of the system; it’s a matter of time, and less time than you think.
I’m not saying that the technology can or should be stopped; that’s simply not realistic. I am saying that, as a society, there are some important conversations we need to start having about what we want things to look like in 10 to 20 years. If we don’t have those discussions, we are going to end up in a world with very high unemployment, where the very few people who hold large capital and those with the STEM skills to design and run the AIs will do very well, while the other 80-90% of us could potentially be unemployable. This is truly scary stuff, McKinsey predicts that by 2025 technology will take over tasks currently performed by hundreds of millions of knowledge workers. This is no longer science fiction.
As humans, our brains evolved to work linearly, and we have a hard time understanding and predicting change that happens exponentially. For example, merely 30 years ago, it was unimaginable that most people would walk around with a device in their pockets that could perform more sophisticated computing than computers at MIT in the 1950s. The huge improvement in power is a result of exponential growth of the kind explained by Moore’s law, which is the prediction that the number of transistors that fit on a chip will double every two years while the chip’s cost stays constant. There is every reason to believe that AI will see similar exponential growth. Just five years ago, the world’s top AI experts at MIT were confident that cars could never drive themselves, and now Google has proven them wrong. Things can advance unimaginably fast when growth becomes exponential.
Some of the most brilliant minds of our times are sounding the alarm bells. Elon Musk said, “I think we should be very careful about AI. If I had to guess, our biggest existential threat is probably that we are summoning the demon.” Stephen Hawking warned, “The development of full-artificial intelligence could spell the end of the human race.”
What was once science fiction is fast becoming a fact of today’s business world. Computers that mimic the human brain are already entering the workforce in the healthcare, financial services and retail sectors, among others.
Like humans, cognitive analytic computers can understand “natural” language (such as English) and learn lessons from the data they analyze, as well as from the users who “mentor” them. In other words, the machines possess an artificial intelligence more powerful than anything seen before.
Unlike humans, cognitive analytic systems can process, analyze and store enormous volumes of data at Internet speed. In addition to tapping conventional databases for the information needed to aid in decision-making, the machines are capable of scanning myriad emails, reports, articles, books and other sources of knowledge to deliver recommendations and reach conclusions beyond the ability of any one person or team of people.
In a 2014 white paper on cognitive analytics, Rajeev Ronanki and David Steier of Deloitte Consulting note that in the healthcare industry, “[cognitive analytic] systems are being used to improve the quality of patient outcomes. A wide range of structured inputs, such as claims records, patient files and outbreak statistics are coupled with unstructured inputs such as medical journals and textbooks, clinician notes and social media feeds. Patient diagnoses can incorporate new medical evidence and individual patient histories, removing economic and geographic constraints that can prevent access to leading medical knowledge.”
In financial services, cognitive analytics is used to recommend and execute trades and to also assist in fraud detection and risk underwriting.
Many of us are familiar with less advanced forms of cognitive analytics. In the consumer electronics realm, examples include Apple’s Siri voice recognition software and the oral command interface used in the Xbox video game system.
Virtual Decision-Making Assistance
It doesn’t take much imagination or intelligence (human or artificial) to envision how cognitive analytics could revolutionize auto insurance, especially the claims sector.
Cognitive analytic computing could be of enormous benefit to an industry that will see fewer claims adjusters in the near future, thanks to the number of veteran adjusters who are retiring or planning to retire. Cognitive analytics could empower the remaining adjusters with decision-making assistance that was previously inconceivable – decision-making based on huge volumes of data drawn from a near-infinite pool of sources.
Not long from now, computers will be able to scan photos of accident damage and instantly retrieve historical data on how similar claims were assessed and settled in the past. For example, a computer could analyze a person’s injuries relative to where they were sitting when the accident occurred and how the injury was sustained.
The systems could also be used in first notice of loss (FNOL). Imagine an intelligent learning system that can reference every text related to previous claims and outcomes, as well as every law and vehicle code from all 50 states, to deliver settlement information in milliseconds.
Let’s say a customer submits an FNOL. “I was in a parking lot, but when I backed out of my space I hit someone driving past.” Based on the information provided, the machine could determine liability and assign fault. It could also decide whether the claim is best processed with the help of a human adjuster or via self-service. If a customer reports an accident that leaves a small scratch on the car and no injuries, the computer would automatically send a self-service text to the claimant’s cell phone so she could take photos of the damage and transmit them back to the computer. The machine would then analyze the photos and develop an assessment.
Yes, the computing system could be that advanced – so advanced that it removes much of the human element from the process.
‘Brain Gain’ Instead of ‘Brain Drain’
Many adjusters in their 50s and 60s are retiring, which means a lot of valuable expertise and experience is leaving the industry. In fact, I’m probably a member of the last generation that remembers widespread use of full-service, multi-skilled adjusters – people who know every aspect of the business. Younger adjusters frequently work in silos. These compartmentalized workers are very skilled in certain things but don’t have the “Renaissance man” backgrounds that allowed their predecessors to wear “multiple hats” when the situations called for it.
Thanks to the new technology, however, the older generation’s experience and know-how doesn’t have to be lost forever. That information and wisdom can be transferred to complex cognitive computing systems that instantly retrieve the data on every one of their past settlements. This will let the remaining adjusters use the machines as virtual assistants, calling on them to provide the most logical settlement paths to the best possible outcomes.
If achieving the best outcomes to claims is the goal, then cognitive computing systems will prove to be an invaluable tool. With access to a virtual universe of prior decision-making (good and bad), cognitive analytics has the potential to help adjusters find the right solution to each and every auto claims case.
In the world of commercial insurance, there exists the very curious role of risk manager. I mean curious in the sense that successful risk managers appear to have superpowers. They are charged with taking the actions necessary to avoid or reduce the consequence of risk across an entire enterprise. Their knowledge must extend deeply into a variety of subjects such as engineering, safety, the subtleties of the business of their employer, insurance (of course), physics, employee motivation and corporate politics and leadership. Their impact can be wide-ranging, from financial (e.g., dollar savings from risk avoidance/mitigation) to personal (the priceless value of the avoidance of employee death or injury).
Sadly, the tyranny of economics restricts the access that businesses have to continuous, high-quality risk management. Full-time risk managers are prevalent in huge, complex, global companies. These firms often self-insure, or purchase loss-sensitive accounts, and the financial value of a risk management position (or department) is clear. The larger mid-market firms can afford to selectively purchase safety consultant services; their insurance broker might perform some of these tasks (especially at renewal), and their insurers may have loss control professionals working some of these accounts. However, for the majority of small businesses, risk management at the professional level is not affordable.
I have toyed with different ideas about how to automate this function to bring the value of a risk manager to the small commercial business segment. My attempts were always unsatisfying. However at a Front End of Innovation conference in Boston, a presentation by Dr. Rafael J. Grossmann (@ZGJR) crystallized the vision. I can now clearly see how existing technology can be combined to create a risk manager avatar.
Dr. Grossmann is a trauma surgeon who practices in Maine. In addition to the normal challenges of his profession, he is one of only four trauma surgeons servicing a very wide area. Although the area is sparsely populated, the challenge of distance and time complicates the delivery of medical services. Dr. Grossmann presented his vision of a medical avatar, a combination of technologies that will perform 80% or more of the routine medical cases in a consistent, timely and cost-effective manner. Combining the technologies of mobile, voice recognition, virtual reality, artificial intelligence, machine learning and augmented reality forms a new silicon entity – a medical doctor avatar. He also introduced a company, sense.ly, that is working to deliver similar services (video here: http://www.sense.ly/index.php/applications/).
If such systems can deliver medical services, then why not risk management? For example, given permission, a system would monitor the purchases of a small company and identify when the historical pattern changes, e.g., when the company begins to buy new types of materials. Using predictive algorithms, the pattern can be compared against others to evaluate if there is likelihood that the company is now performing new business operations. The avatar could then contact the small business, or could signal human intervention by an underwriter to evaluate the necessity for an endorsement to a policy to cover the new business operation. Eventually, some of these interventions would also be handled through machine-to-machine communication and would allow the endorsement to take place automatically.
Someone will build a risk management avatar. The question is, who will do it first?
The robots are here. Not the humanoid versions that you see in Hollywood movies, but the invisible ones that are the brains behind what look like normal online front-ends. They can educate you, advise you, execute trades for you, manage your portfolio and even earn some extra dollars for you by doing tax-loss harvesting every day. These robo-advisers also are not just for do-it-yourself or self-directed consumers; they’re also for financial advisers, who can offload some of their more mundane tasks on the robo-advisers. This can enable advisers to focus more on interacting with clients, understanding their needs and acting as a trusted partner in their investment decisions.
It’s no wonder that venture capital money is flowing into robo-advising (also called digital wealth management, a less emotionally weighted term). Venture capitalists have invested nearly $500 million in robo-advice start-ups, including almost $290 million in 2014 alone. Many of these companies are currently valued at 25 times revenue, with leading companies commanding valuations of $500 million or more. This has motivated traditional asset managers to create their own digital wealth management solutions or establish strategic partnerships with start-ups. Digital wealth management client assets, from both start-ups and traditional players, are projected to grow from $16 billion in 2014 to roughly $60 billion by end of 2015, and $255 billion within the next five years. However, this is still a small sum considering U.S. retail asset management assets total $15 trillion and U.S. retirement assets total $24 trillion.
What has caused this recent “gold rush” in robo-advice? Is it just another fad that will pass quickly, or will it seriously change the financial advice and wealth management landscape? To arrive at an answer, let’s look at some of the key demographic, economic and technological drivers that have been at play over the past decade.
The need for digital wealth management and the urgent need to combine low-cost digital advice with face-to-face human advice have arisen in three primary market segments, which many robo-advisers are targeting:
Millennials and Gen Xers: More than 78 million Americans are Millennials (those born between 1982 and 2000), and 61 million are Gen Xers (those born between 1965 and 1981); accordingly, this segment’s influence is significant. These groups demand transparency, simplicity and speed in their interactions with financial advisers and financial services providers. As a result, they are likely to use online, mobile and social channels for interactive education and advice. That said, a significant number of them are new to financial planning and financial products, which means they need at least some human interaction.
Baby Boomers: Baby boomers, numbering 80 million, are still the largest consumer segment and have retail investments and retirement assets of $39 trillion. Considering that this segment is either at or near retirement age, the urgency to plan for their retirement as well as draw down a guaranteed income during it is critical. The complexity of planning and executing this plan typically goes beyond what today’s automated technologies can provide.
Mass-Affluent & Mass-Market: Financial planning and advice has largely been aimed at high-net-worth (top 5%) individuals. Targeting mass-affluent (the next 15%) and mass-market (the next 50%) customers at an affordable price point has proven difficult. Combining automated online advice with the pooled human advice that some of the digital wealth management players offer can provide some middle ground.
Technical advances have accompanied demographic developments. The availability of new sources and large volumes of data (i.e., big data) has meant that new techniques are now available (see “What comes after predictive analytics?”) to understand consumer behaviors, look for behavioral patterns and better match investment portfolios to customer needs.
Data Availability: The availability of data, including personally identifiable customer transactional level data and aggregated and personally non-identifiable data, has been increasing over the past five years. In addition, a number of federal, state and local government bodies have been making more socio-demographic, financial, health and other data more easily available through open government initiatives. A host of other established credit and market data companies, as well as new entrants offering proprietary personally non-identifiable data on a subscription basis, complement these data sources. If all this structured data is not sufficient, one can mine a wealth of social data on what customers are sharing on social media and learn about their needs, concerns and life events.
Machine Learning & Predictive Modeling: Techniques for extracting insights from large volumes of data also have been improving significantly. Machine learning techniques can be used to build predictive models to determine financial needs, product preferences and customer interaction modes by analyzing large volumes of socio-demographic, behavioral and transactional data. Big data and cloud technologies facilitate effective use of this combination of large volumes of structured and unstructured data. In particular, big data technologies enable distributed analysis of large volumes of data that generates insights in batch-mode or in real-time. Availability of memory and computing power in the cloud allows start-up companies to scale on demand instead of spending precious venture capital dollars setting up an IT infrastructure.
Agent-Based Modeling: Financial advice; investing for the short-, medium- and long-term; portfolio optimization; and risk management under different economic and market conditions are complex and interdependent activities that require years of experience and extensive knowledge of numerous products. Moreover, agents have to cope with the fact that individuals often make investment decisions for emotional and social reasons, not just rational ones.
Behavioral finance takes into account the many factors that influence how individuals really make decisions, and human advisers are naturally skeptical that robo-advisers will be able to match their skills interpreting and reacting to human behavior. While this will continue to be true for the foreseeable future, the gap is narrowing between an average adviser and a robo-adviser that models human behavior and can run scenarios based on a variety of economic, market or individual shocks. Agent-based models are being built and piloted today that can model individual consumer behavior, analyze the cradle-to-grave income/expenses and assets/liabilities of individuals and households, model economic and return conditions over the past century and simulate individual health shocks (e.g., need for assisted living care). These models are assisting both self-directed investors who interact with robo-advisers and also human advisers.
Evolution of Robo-advisers
We see the evolution of robo-advisers taking place in three overlapping phases. In each phase, the sophistication of advice and its adoption increases.
First Generation or Standalone Robo-Advisers: The first generation of robo-advisers targets self-directed end consumers. They are standalone tools that allow investors to a) aggregate their financial data from multiple financial service providers (e.g., banks, savings, retirement, brokerage), b) provide a unified view of their portfolio, c) obtain financial advice, d) determine portfolio optimization based on life stages and e) execute trades when appropriate. These robo-advisers are relatively simple from an analytical perspective and make use of classic segmentation and portfolio optimization techniques.
Second Generation or Integrated Robo-Advisers: The second generation of robo-advisers is targeting both end consumers and advisers. The robo-advisers are also able to integrate with institutional systems as “white labeled” (i.e., unbranded) adviser tools that offer three-way interaction among investors, advisers and asset managers. These online platforms are variations of the “wrap” platforms that are quite common in Australia and the UK, and offer a cost-effective way for advisers and asset managers to target mass-market and even mass-affluent consumers. In 2014, some of the leading robo-advisers started “white labeling” their solutions for independent advisers and linking with large institutional managers. Some larger traditional asset managers also have started offering automated advice by either creating their own solutions or by partnering with start-ups.
Third Generation or Cognitive Robo-Advisers: Advances in artificial intelligence (AI) based techniques (e.g., agent-based modeling and cognitive computing) will see second generation robo-advisers adding more sophisticated capability. They will move from offering personal financial management and investment management advice to offering holistic, cradle-to-grave financial planning advice. Combining external data and social data to create “someone like you” personas; inferring investment behaviors and risk preferences using machine learning; modeling individual decisions using agent-based modeling; and running future scenarios based on economic, market or individual shocks has the promise of adding significant value to existing adviser-client conversations.
One could argue that, with the increasing sophistication of robo-advisers, human advisers will eventually disappear. However, we don’t believe this is likely to happen anytime in the next couple of decades. There will continue to be consumers (notably high-worth individuals with complex financial needs) who seek human advice and rely on others to affect their decisions, even if doing so is more expensive than using an automated system. Because of greater overall reliance on automated advice, human advisers will be able to focus much more of their attention on human interaction and building trust with these types of clients.
Implications to Financial Service Providers
How should existing producers and intermediaries react to robo-advisers? Should they embrace these newer technologies or resist them?
Asset Managers & Product Manufacturers: Large asset managers and product manufacturers who are keen on expanding shelf-space for their products should view robo-advisers as an additional channel to acquire specific type of customers – typically the self-directed and online-savvy segments, as well as the emerging high-net-worth segment. They also should view robo-advisers as a platform to offer their products to mass-market customers in a cost-effective manner.
Broker Dealers and Investment Advisory Firms: Large firms with independent broker-dealers or financial advisers need to seriously consider enabling their distribution with some of the advanced tools that robo-advisers offer. If they do not, then these channels are likely to see a steady movement of assets – especially of certain segments (e.g., the emerging affluent and online-savvy) – from them to robo-advisers.
Registered Independent Advisers and Independent Planners: This is the group that faces the greatest existential threat from robo-advisers. While it may be easy for them to resist and denounce robo-advisers in the short term, it is in their long-term interest to embrace new technologies and use them to their advantage. By outsourcing the mechanics of financial and investment management to robo-advisers, they can start devoting more time to interacting with the clients who want human interaction and thereby build deeper relationships with existing clients.
Insurance Providers and Insurance Agents: Insurance products and the agents who sell them also will feel the effects of robo-advisers. The complexity of many products and related fees/commissions will become more transparent as the migration to robo-adviser platforms gathers pace. This will put greater pressure on insurers and agents to simplify and package their solutions and reduce their fees or commissions. If this group does not adopt more automated advice solutions, then it likely will lose its appeal to attractive customer segments (e.g., emerging affluent and online-savvy segments) for whom their products could be beneficial.
Product manufacturers, distributors, and independent advisers who ignore the advent of robo-advisers do so at their own risk. While there may be some present-day hype and irrational exuberance about robo-advisers, the long-term trend toward greater automation and integration of automation with face-to-face advice is undeniable. This situation is not too dissimilar to automated tax-advice and e-filing. When the first automated tax packages came out in the ’90s, some industry observers predicted the end of tax consultants. While a significant number of taxpayers did shift to self-prepared tax filing, there is still a substantial number of consumers who rely on tax professionals to file their taxes. Nearly 118 million of the 137 million tax returns in 2014 were e-filings (i.e., electronically filed tax returns), but tax consultants filed many of them. A similar scenario for e-advice is likely: a substantial portion of assets will be e-advised and e-administered in the next five to 10n years, as both advisers and self-directed investors shift to using robo-advisers.