Long before the COVID-19 pandemic, insurers were investing in digital transformation, spurred by the rise of startups. Those investments took on new urgency as the pandemic forced businesses across industries to move to digital operations to stay afloat.
Over the long term, no technology will prove as vital to insurers’ agility and success as artificial intelligence, whose far-reaching impact will define the next wave of insurtech innovation.
Legacy players and nascent startups alike will leverage AI and machine learning to enhance customer service, speed claims processing and improve the accuracy of underwriting – enabling insurers to match customers to the right products, operate with greater efficiency and achieve better results.
Though insurance is often cast as slow to embrace technology and innovation, in a certain respect AI is very much within the industry’s wheelhouse. Since the first actuaries began their work in the 17th century, insurance has relied heavily on data – and as AI empowers insurers to do even more with vast swaths of data, the benefits will redound to providers and policyholders alike.
Bringing Customer Service to the Next Level
In today’s digital economy, personalization is all the rage. Customers crave tailored, relevant experiences, offers and promotions that reflect their unique backgrounds, needs and interests – and they increasingly expect businesses to deliver these experiences as a basic standard of service.
While personalization is often discussed in the context of sectors like e-commerce, the insurance industry is no exception to this trend. According to an Accenture survey, 80% of customers expect their insurance providers to customize offers, pricing and recommendations.
Of course, delivering bespoke experiences requires an abundance of customer data – and customers are more than willing to provide it in exchange for personalized service; 77% told Accenture that they’d share their data to receive lower premiums, quicker claims settlement or better coverage recommendations.
Because personalization can only deliver on its promise if it’s holistic and omnichannel, the most successful insurers will be those that don’t view personalized engagements as one-offs – a tailored email here, a promotion there – but that consistently provide personalization at every stage of the customer journey.
What will that look like in practice? AI chatbots will become a lot more “chat” and a lot less “bot,” not only providing 24/7 customer service but also using cutting-edge methods like natural language processing (NLP) to better understand what customers actually need and to conduct more natural, intuitive conversations. Underwriting will become much more precise as machines crunch massive sets of data – reams of usage and behavioral data generated by customers and their IoT devices, as well as relevant geographic, historic and other information – to create customized policies that reflect a policyholder’s true level of risk.
Harnessing the power of AI, insurers can also streamline claims processing as part of a comprehensive digital strategy. Forward-thinking providers will increasingly integrate automated customer service apps into their offerings. These apps will handle most policyholder interactions through voice and text, directly following self-learning scripts that will be designed to interface with the claims, fraud, medical service and policy systems.
As a McKinsey analysis noted, with automated claims processing, the turnaround time for settlement and claims resolution will start to be measured in minutes rather than days or weeks. Meanwhile, human claims management associates will be free to shift their focus to more complicated claims, where their insights, experience and expertise are truly needed.
These transformative applications of AI will unlock revenue opportunities, improve risk management and help insurers deliver a new level of personalized customer service. But if AI will act as the great enabler, what will enable AI itself?
The answer lies in a robust digital core, which is vital to facilitating efficient business processes, maintaining resilience in an unpredictable world and supporting the rollout of new products and business offerings. Whether insurers manage to achieve that kind of digital agility will determine their ability to survive and thrive in a landscape that’s shifting faster than ever.
Technologies like machine learning, the Internet of Things (IoT), robotic process automation (RPA) and natural language processing (NLP) were already hot topics in P&C insurance before the world was turned upside down in 2020 due to the pandemic. These and many other “transformational” technologies have great potential for insurers in the rethinking and optimization of distribution, underwriting, claims and many other parts of the business. So, it is important to ask the question – how have the initiatives that leverage these technologies changed due to the pandemic?
Are personal and commercial lines carriers still moving forward with projects in 2021? Do executives still have the same expectations about the potential of these technologies to transform their business?
We answer these questions in detail for 13 specific technologies in two new SMA research reports, one covering personal lines and the other covering commercial lines.
However, I won’t leave you hanging in this blog, wondering about the answers to those questions. The short answer is yes – P&C insurers generally plan to move forward in 2021 with projects that leverage various technologies that have the power to deliver significant results and competitive advantage. The technologies we follow closely and have profiled in our reports have been organized into three strategic planning horizons: short-term, near-term and long-term.
For both personal and commercial lines, technologies in the AI family play heavily in the short-term category. Machine learning, NLP, RPA, computer vision and new user interaction technologies all rank high in terms of their potential to transform and in the level of activity underway or planned by insurers. Technologies that fall into the near-term or long-term horizons include wearables, blockchain, voice, AR/VR (augmented reality/virtual reality), 5G and autonomous vehicles. All have potential in insurance and will likely be incorporated into projects by innovators over the next couple of years but will not make it into broad, mainstream application until midway to late in the decade.
Our research on transformational technologies, when viewed in concert with our SMA Market Pulse surveys, shows that in some cases proofs of concept (POCs) and new projects have been put on hold in 2020, but all indications point to full steam ahead in 2021. Major projects already underway are continuing, and insurers state that they do not want to lose momentum for foundational projects like core systems. Projects that include transformational technologies needed to address digital gaps that were exposed during the pandemic have been raised in priority.
In many ways, the pandemic is accelerating digital transformation across all industries, including insurance. Transformational technologies will play an outsized role in that transformation and look to be important components of insurers’ plans for 2021 and beyond.
Insurance company executives are being pressured by board discussions, distribution channel partners and customer service requirements to more aggressively leverage the “shiny objects” that insurtech offers. Artificial intelligence (AI) is one of insurtech’s brightest contributions, and it seems natural for insurers to use advances in AI — including machine learning (ML), natural language processing (NLP) and robotic process automation (RPA) — to leapfrog competitors.
Unfortunately, not every insurer is ready for AI or able to take full advantage of the opportunities in this category of emerging technologies. There are, however, several ways insurers can prepare and evolve to a position of strength from which AI can make a strategic business impact.
1. Ditch Dirty Data
For a variety of reasons, insurers tend to have a good amount of “dirty” data, rife with inconsistent formats or standards, incomplete conversions resulting from merger and acquisition activity and data transfer from paper files. A proliferation of dirty data can put insurers in the untenable position of sacrificing whatever valuable intelligence may exist in historical files to a “Day Forward” strategy.
Insurers looking to prioritize AI projects must invest in cleansing bad data and improving data mastery. Those efforts will naturally include improving access to, and use of, both structured and unstructured data. The “magic” of AI gives the impression this technology is a silver bullet capable of maximizing the value of the unstructured data prevalent in handwritten forms, PDFs, images, email and text messages and social data, which increasingly inundate insurer workflows. However, the organization of clean and available data is a precursor to AI implementation.
A recent report by Eric Weisberg and Mitch Wein of Novarica, “MDM in Insurance: Expansion and Key Issues,” details the need for insurers to invest more heavily in improving data mastery and hiring for positions such as chief data officer or data scientist, instead of purely tech talent. “Insurers are placing a priority on data initiatives to support their predictive modeling and AI programs,” Weisberg and Wein wrote. “High data quality is imperative for digitization where data is being exposed to outside parties. Existing and emerging data regulations are also driving a need for improved data governance. Chief data officers and multi-tiered data governance organizations are becoming more prevalent as data is increasingly being treated as an asset. Challenges exist with organization, resourcing, process and funding that can stymie the results of well-intentioned data programs.”
2. Cultivate the Right Culture
The fast-evolving nature of technology often means insurers are in a fluid state of decision-making about deployment. As innovation further penetrates traditional industry settings and transforms basic processes and products, insurers must decide if the organization’s culture and leadership are truly capable of committing to the journey of transformation, let alone arriving at the destination.
New tools, such as AI solutions, will demand new skills of managers who have built careers leading and inspiring people, and who understand the importance of change management to the organization. So, sorting out the boardroom and operational priorities of the CFO and CIO, or the VP of IT and COO, can help ensure solid business cases and implementation strategy for innovation — such as AI initiatives.
3. Prioritize the Policyholder
In addition to cleansing dirty data and strengthening internal change management, preparing to better leverage AI should include a re-prioritization around the policyholder and the customer experience. Insurers need more customer-centric processes from the ground up and a reinvention of existing products and processes that treat the policy as an attribute of the customer instead of the other way around. Customer acquisition is notoriously expensive, and insurers face the additional challenge of relating an age-old industry and product to a new generation of consumers. To be successful, the gap between old and new, and between company and customer, must be narrowed substantially.
AI can aid such efforts through innovations such as natural language processing (NLP), which recognize information included in voice conversations or recordings and then quickly and accurately deliver relevant policy files or information. Chatbots can also improve the speed of customer service interactions, and ultimately the speed at which policyholder concerns are resolved. Claims service is good example of a process in which insurers are already starting to see the benefits of incorporating AI solutions, and are using this technology to do everything from reporting first notice of loss (FNOL) to initiating claim processing, or even deploying an adjuster, if necessary.
As insurers prioritize spending on AI initiatives and implementations, the danger is ignoring persistent shortfalls in important areas — such as data mastery, operations and even underwriting. And, it is important to recognize that innovation implemented in the form of an AI solution alone is not, and never will be, a viable strategy. AI can be an enabler of a strategy. But without clearly defined goals and a flexible operating model capable of supporting an evolving and demanding policyholder portfolio, even a successful AI implementation can end up as no more than a footnote.
Process automation, machine learning and other types of AI initiatives will continue to make for compelling business cases. To realize full potential and benefits, those tools should be focused on winning clients and implementing accessible, 24/7 customer service and operationally optimizating to support competitive differentiation. Cost savings from AI will typically flow as a by-product. But, without leadership, champions who embrace and drive change and organizational data mastery, the AI tools will be underused and unlikely to fulfill the promise of growth and service excellence.
Insurers have a near-constant stream of unstructured data at their disposal that can be used to drive growth by improving policyholder retention and identifying cross-sell and upsell opportunities. One of the challenges for insurers is sorting through this mountain of unstructured data quickly to gain an accurate understanding of the sentiment of their customers in real time.
The last few years have seen sentiment analytics become a critical component of customer feedback strategies for companies of all sizes. Sentiment analytics uses a combination of natural language processing (NLP), machine learning (ML) and deep text analytics to bring out the nuances hidden in the text.
Sentiment is easier to translate and analyze than it is to express. Sentiment analytics, also referred to as opinion mining, is a technique to abstract the underlying sentiment from textual data. Usually, this customer feedback is unstructured data flowing in from multiple channels, such as:
Call center logs
Social media posts
The idea is to understand not only the nature of the feedback but also to derive context out of it. However, sentiments are complicated. Analyzing sentiments, even more so.
Domain-Specific Sentiment Analytics
The complexity of spoken language makes it difficult if not impossible to derive sentiment accurately every time. Teaching a machine to understand such things as tonality, cultural lingo and slang, or the ability to discount grammatical errors, and comprehend rhetoric such as irony or sarcasm are all difficult at present. Existing sentiment analytics tools are not equipped to identify the true sentiment of these types of dialogues.
Although sarcasm is a problematic form of language to detect, there are other complex statements that machines are learning to comprehend. Consider the following statement: “Rocketz Insurance Company has always offered me great pricing, but at times I have not been happy with their response time for questions about my policy.” This sentence has a negative as well as a positive connotation, and sentiment analytics come into play. The first part of the sentence can be identified as a positive feeling, and the other half is identified as negative.
Sentiment analytics specific to industry lines play a key role. The accuracy of identifying the sentiment of data can be increased by training the system (machine learning) on a specific domain, such as the insurance industry. For example, the terms “garaging” and “towing” have a greater meaning in the insurance industry as opposed to, for example, manufacturing. Therefore, if a client makes a comment about either, it would have more meaning for insurance than other industries.
In the insurance industry, sentiment analytics can be used in a multitude of ways that directly affect business, such as:
Improving retention rates
Identifying cross-sell/upsell opportunities
Improving Retention Rates
Having the ability to quickly and easily identify the sentiment of policyholders whose auto policy will renew in 60 days or less is a good example. Let’s say we have 1,000 auto policies that are up for renewal by the end of 2019, and the priority for the renewal team is to contact policyholders who are “detractors.” The sentiment of their conversations and interactions, regardless of the channel, has a low score. The challenge for the renewal team is: Who do they contact first? Are all “detractors” equal in their dissatisfaction with their auto policy? And what may be the root cause or causes of their dissatisfaction?
The renewal team can target these policyholders with a strategy to retain these policyholders with insight on why the policyholder may be unhappy before contacting them. Maybe it was a bad claim experience, or they are unhappy with their premium and started shopping elsewhere. It’s not enough to simply understand who a “detractor” is; you need to understand why.
Identifying Cross-sell/Upsell Opportunities
Using the example of our 1,000 auto policies that are up for renewal, what about the policyholders who are “promoters” and happy with their auto coverage? This is an ideal time for the renewal team to contact these policyholders and thank them for their business as a minimum. But is there a cross-sell or upsell opportunity here? For the renewal team focusing on this segment of policyholders, it would be helpful to have some idea why their sentiment is high before contacting them with a potential offer.
In addition to identifying immediate opportunities that can be acted on, sentiment analytics can help insurers understand trends such as:
The sudden demand for a product type
The like or dislike of a specific customer experience
Many homeowners are considering cyber insurance to protect themselves from identity theft and invasion of privacy. Sentiment analytics can provide insurers who currently do not offer cyber insurance a heads-up that maybe they should consider offering cyber insurance as a cross-sell/upsell opportunity.
Everyone wants to identify and correct bad customer experiences, especially a bad claims experience. What about good customer experiences? Going back to our auto policy example, something as simple as having the ability to easily obtain an auto insurance card online, or easily reach a customer service representative. can be a positive customer experience that sales and marketing may want to promote.
Sentiment analytics can help insurers sort through a continuous stream of unstructured data to identify opportunities for increasing revenue and identifying trends. Currently, sentiment analytics is not perfect, but focusing on a specific domain such as the insurance industry will increase accuracy.
Sentiment analytics can be a powerful tool if leveraged starting at the earliest touch point, even if it begins with a small set of customers.
Artificial intelligence is the new electricity. We hear it will fundamentally shift the balance of power between labor and capital, mostly by rendering labor obsolete. It will enable and empower transformative technologies that will rearrange the sociopolitical landscape and may lead to humanity’s transcendence (or extinction) within our lifetimes. As it changes the world, it will necessarily rewrite the rules of insurance. That’s the myth, and the nature of the headlines.
Interestingly, insurance is heavy on intellectual property (think of proprietary underwriting models), technology and data. And AI is hungry; hungry for data, of course, but also hungry for systems that can be automated and for proprietary classification problems that can be improved. That places insurance right in the appetite of artificial intelligence and its promise of transformation. If we want to act on artificial intelligence’s transformational potential, we need to understand what it actually is, separate the technologies from the hype and develop a practical understanding of what is required to implement AI-powered solutions in the insurance sector. This article will highlight these three steps and offers a realistic approach for carriers to take advantage of the opportunities.
Defining Artificial Intelligence
Unfortunately, our first step is also our hardest, as a working definition of artificial intelligence is difficult. The scope of the term AI is broad, and it requires careful consideration to avoid becoming hopelessly confounded with its own hype. It is also challenging to come to a clear definition of natural intelligence, which leaves us struggling for a definition of artificial intelligence because the latter is so often compared to the former.
AI tends to be discussed in two flavors. The first is general artificial intelligence (also, artificial general intelligence and strong AI). GAI is machinery capable of human-level cognition, including a general problem-solving capability that is potentially self-directed and broadly applicable to many kinds of problems. GAI references are accessible through fictional works, such as C-3PO in Star Wars or Disney’s eponymous WALL-E. The most important feature of GAI is that it does not currently exist, and there is deep debate about its potential to ever exist.
The second is usually referred to as narrow AI. Narrow AI is task-specific and non-generalizable. Examples include facial recognition on Apple’s iPhone X and speech-to-text transliteration by Amazon’s Alexa. Narrow AI looks and feels a lot like software or, perhaps, predictive models. Narrow AI can be described as a class of modeling techniques that fall under the category of machine learning.
What is machine learning? Imagine a set of input data; this data has one or more potential features of interest. Machine learning is a technique for mapping the features of input data to a useful output. It is characterized by statistical inference, as advanced techniques often underlie machine learning predictive models. Through statistical modeling, software can infer a likely output given a set of input features. The predictive accuracy of machine learning methods increase as their training data sets increase in size. As the machine ingests more data, it is said to learn from that data. Hence, machine learning.
Perhaps most important of all, machine learning (as an implementation of narrow AI) is real and here today; for the remainder of our discussion when we say “AI,” we mean narrow AI or machine learning.
Beyond the Hype
The hype around AI and its potential is extensive. Silicon Valley billionaires opine on the potential implications of the technology, including comparing its power to nuclear weapons. Articles endlessly debate if and how quickly AI will structurally unemploy vast swaths of white collar workers. MIT’s Technology Review provides a nice summary of the literature, stating that up to half of all jobs worldwide could be eliminated in the next few decades.
AI may well have this kind of impact. And the social, political and economic implications of that impact, especially around questions of potential large-scale unemployment, deserve careful long-term consideration. However, executives and business owners need to evaluate technology investments today to improve their current competitive position. From that perspective, we find it more practical to focus on examining which existing tasks could be automated by AI today.
In 2012, researchers trained pigeons to recognize people based only on their faces as part of a study on cognition. Suppose you had millions of face-recognizing pigeons; this force of labor could be deployed in a comprehensive facial recognition system — a system remarkably similar in function to the facial recognition AI of devices like modern smart phones. It turns out pigeons have also been trained to recognize voices, spot cancers on X-rays and count, among a host of other tasks related to headline-grabbing AI achievements.
The metaphor is admittedly silly. Instead of pigeons, imagine an army of virtual robots capable of classifying information from the real world to produce a machine-readable data set. In machine learning language, these robots take unstructured data and make it structured. Said robots resemble the automation machinery of a factory; like spot welders tirelessly joining steel members to form automobile frames, our virtual robots tirelessly recognize if a face is featured in a photograph. In contemplating the question, what could be automated with AI, a useful starting place is the army of robots (or pigeons!). For example:
What existing analyses could be improved or optimized? Could pricing or underwriting be improved using better classifiers or non-linear modeling approaches?
What data currently exist at the firm that could be made available for new types of analysis? Claims adjusters’ notes can be processed by natural language algorithms and cross-referenced with photos of physical damage or prior inspections.
What data would you analyze if it could be made available? What if you could listen to all the policyholder calls received by your customer service department and annotate which questions stumped the customer service representatives? Or which responses lead to irritation in the policyholders’ voices?
Bringing AI to Insurance
What is an insurer to do? Start by not fretting. We propose two considerations to facilitate a sleep-at-night perspective. First, insurers are already good at AI or its precursor technologies. The applicability of AI in the present and near future is entirely based on narrow AI technologies. For example, natural language processing and image recognition are both machine learning implementations with working business applications right now. Both use predictive models to achieve results. The software may be artificial neural networks trained on vast data sets, but they are nonetheless conceptually compatible with things insurance carriers have used for years, like actuarial pricing models. The point is that the application of AI is an incremental step forward in the types of models and data already applied in the business.
Second, sorting through the hype requires a staple of good business decision making: the risk-cost-benefit analysis. Determining which technologies are worth investment is within scope for decision makers that otherwise know how to make selective investments in growing the capabilities of their firm. The problems faced by a carrier are much bigger than sorting out AI if management lacks the basic skillset for making business investments.
Providing an inventory of every application of AI is beyond the scope of this article. DeepIndex provides a list of 405 at deepindex.org, from playing the Atari 2600 to spotting forged artworks. Instead, suppose that AI, like electricity, will be broadly applicable across industries and functions, including the components of the insurance value chain from distribution to pricing and underwriting to claims. The goal is to identify and implement the AI-empowered solutions that will further a competitive advantage. Our view is that carriers’ success with AI requires three key ingredients: data, infrastructure and talent.
Data: AI might be considered the key that unlocks the door of big data. Many of the modeling techniques that fall under the AI umbrella are classification algorithms that are data hungry. Unlocking the power of these methods requires sufficient volume of training data. Data takes several forms. First, there are third party data sources that are considered external to the insurance industry. Aerial imagery (and the processing thereof) to determine building characteristics or estimate post-catastrophe claims potential are easy examples. Same with the vast quantities of behavioral data built on the interactions of users with digital platforms like social media and web search. Closer to home, insurance has long been an industry of data, and carriers are presumed to have meaningful datasets in claims, applications and marketing, among others.
Infrastructure: Accessing the data to feed the AI requires a working infrastructure. How successfully can you ingest external data sources? How disparate and unstructured can those sources be? Cloud computing is not necessarily a prerequisite to successful AI, but access to vast, scalable infrastructure is enabling. Are your information systems equipped, including security vetting, to do modeling in the cloud? Can you extract your internal data into forms that are ready to be processed using advanced modeling techniques? Or are you running siloed legacy systems that prevent using your proprietary data in novel ways?
Talent: Add data science to the list of AI-related buzzwords. We claimed earlier that many of the advancements attributed to narrow AI are predictive models conceptually like modeling techniques already used in the insurance industry. However, the fact that your pricing actuary conceptually appreciates an artificial neural net built for fraud detection using behavioral data does not mean you have the in-house expertise to build such a model. Investments in recruiting, training and retaining the right talent will provide two clear benefits. The first benefit is being better equipped to do the risk-cost-benefit analysis of which data and methods to explore. The second is having the ability to test and, ultimately, implement.
In Aon’s 2017 Global Insurance Market Outlook we explored the idea of the third wave of innovation as propounded by Steve Case, founder of AOL, in his book, “The Third Wave: An Entrepreneur’s Vision of the Future.” The upshot of the third wave for insurers was that partnership with technology innovators, rather than disruption by them, would be the norm. This approach applies now more than ever as technological innovators continue to unlock the potential of AI. If you don’t have the data, or the infrastructure, or the talent to bring the newest technologies to bear, you can partner with someone that does. Artificial intelligence is real. While the definitions are somewhat vague – is it software, predictive models, neural nets or machine learning – and the hype can be difficult to look past, the impacts are already being felt in the form of chatbots, image processing and behavioral prediction algorithms, among many others. The carriers that can best take advantage of the opportunities will be those that have a pragmatic ability to evaluate tangible AI solutions that are incremental to existing parts of their value chain.
Maybe true, but that does not make it daunting. The core of insurance is this: Hire the right people, give them the infrastructure they need to evaluate risk better than the competition and curate the necessary data to feed the classification models they build. AI hasn’t, and won’t, change that.