The global market for blockchain in insurance is expected to grow more than 50% annually through 2031 because of tangible benefits such as added security and transaction transparency. These are possible due to blockchain’s distributed architecture, which allows users to exchange information confidently without third-party involvement. 

Let’s go through some prominent use cases of blockchain solutions for insurance to see how insurance companies can benefit. 

Anti-money laundering

AML regulations require financial and insurance institutions to have specific mechanisms to detect and prevent money laundering. Thus, organizations set up robust customer due diligence (CDD) and know-your-customer (KYC) processes to assess potential customers before providing any services to them. Verifying the customer’s identity entails gathering information about the client's background, source of funds and transaction patterns. While the KYC and CDD mechanisms have proved beneficial to preventing money laundering, they require substantial investments, as financial institutions must collect, enter and process vast amounts of customer data once a client engages with them.  

Private blockchains can help insurance companies save hours of work collecting and verifying customer data and, thus, save thousands of dollars by acting as a shared storage of customer information for financial institutions. Whenever customer information is updated, a new block is added to the chain, which maintains all transactions and serves as a trusted and traceable source of client information. 

When a customer needs to engage with another financial or insurance organization, this organization can request access to the customer’s data stored in the blockchain. This way, companies streamline customer onboarding and eliminate data inconsistency across multiple institutions.

See also: Blockchain's Future in Surety Industry

Fraud mitigation

Insurance companies can use blockchain technology to combat fraud. Because blockchain presents an immutable ledger that records transactions in a secure and tamper-proof manner, once data is recorded on a blockchain, it cannot be altered or deleted. This way, insurance companies get an auditable trail of all transactions, which allows them to identify discrepancies, patterns or anomalies indicative of fraudulent activity, such as multiple claims for the same incident.

As a result, insurance businesses strengthen trust between the company and its customers: Insurance providers get a powerful tool to verify claims authenticity while customers gain more transparency in claims settlement. 

Smart contracts

Blockchain is a backbone technology in smart contracts, which can automate the claim settlement process in insurance. Smart contracts are designed to execute actions based on contract terms automatically once predefined conditions are met, without third-party involvement. Here are some examples of how smart contracts contribute to parametric and peer-to-peer insurance efficiency.   

Parametric insurance

With parametric or index-based insurance, policyholders get their payout based on the occurrence of a trigger event rather than the damage incurred by the event. The occurrence of the trigger event is detected by oracles, which are reliable sources of information. Once the trigger event occurs, smart contracts automatically trigger the payout to the policyholder without manual intervention. 

Peer-to-peer insurance

Peer-to-peer insurance (P2P) is an insurance model where individuals (family members, relatives or business partners) consolidate their resources to insure each group member against specific incidents. These individuals can encode P2P insurance terms and conditions into smart contracts that will be automatically executed once one of those individuals files a claim.

Data privacy

Keeping data secure, accurate and integral in the insurance industry is paramount. Serving as a future-proof ledger for all transactions and data entries, blockchain can help ensure data privacy, reliability and accuracy. For instance, insurance companies can use blockchain technology to store the data collected with oracles for parametric insurance, providing for the integrity and transparency of the data used to determine policyholder eligibility for payout. Similarly, blockchain can help ensure transparency and joint trust among the P2P insurance members.

Reinsurance

Reinsurance provides financial protection to insurance companies in case of large payouts. Blockchain can make reinsurance more transparent and efficient, fostering trust between reinsurance and ceding companies. With self-executing contracts, both parties can benefit from faster and more transparent claim settlement. Additionally, partners can securely exchange data, eliminating fraudulent activities and improving the accuracy of reinsurance processes.

See also: How to Become a Future-Ready Insurer

A real-life example

One of the most prominent examples of how insurance companies embrace smart contracts on the blockchain is the insurtech company Lemonade. Together with renowned companies such as Avalanche, Chainlin, and Etherics, Lemonade launched a blockchain-powered crop protection insurance program. 

As farmers' harvests in rural Africa heavily rely on the weather, insuring their crops against adverse weather is a reasonable step. Still, traditional indemnity insurance is not financially viable for African regions: Farmers can’t afford high premiums, including claims handling expenses and service coverage. As a solution, Lemonade worked out a parametric crop insurance program hosted on blockchain, which allows for triggering smart contract functions once drought is detected in the region. The program’s cost-effectiveness and transparent claim settlement process make the premiums affordable to African farmers and help them through tough seasons.

Summing up

Blockchain has found its niche in the insurance industry, and it is uniquely positioned to benefit from the technology. Blockchain helps insurance companies automate various effort-intensive processes related to claim management and increase mutual trust among the parties involved in the insurance process. Still, to embrace all these multiple benefits, companies must carefully plan blockchain implementation and ensure its seamless integration into their IT environment and business processes. Do not hesitate to partner with blockchain professionals if you lack the required expertise. Professional consultants can help you reduce costs for technology implementation and significantly accelerate the process.

While the move to driverless cars seems to have hit the pause button, two major companies are planning to put fully autonomous big rigs on highways by the end of the year and have thousands on U.S. roads not too many years after.

That will create lots of nuances for insurers to navigate as the responsibility for accidents shifts from drivers and fleet owners to the developers and operators of the AI managing the semis. You can get a peek into the complexities in store by looking at the lawsuits Tesla has faced over whether its Full Self-Driving AI or the driver is responsible for accidents, including ones that have injured or even killed people. Other sorts of issues will crop up, too, likely including theft, given that there will no longer be a person minding the truck and its contents en route. 

And those are just the beginning of the effects that autonomous trucks will have as they rewire the world's supply chains. Trucks already transport some 12 billion tons of freight every year just in the U.S., according to the American Trucking Associations, and autonomous trucks will carry far more freight, and faster. At the moment, truck drivers are only allowed to operate their big rigs 11 hours a day in the U.S., but autonomous vehicles never get tired and never have to stop, except to refuel. Multi-day trips will take less than half the time they do now. So trucks will absorb business from other modes of transportation, especially planes. The trucking industry says it has a shortage of about 80,000 drivers in the U.S. No more. Not once autonomous trucks really get going. 

Truck stops will change--we'll need more people to pump gas for these driverless trucks... but won't need so many showers or so much coffee. Warehouses will be relocated and managed differently, given that the speed of shipping will rise while the costs drop. Traffic patterns will change as many more trucks pile onto roads. (I can hardly wait to see how much worse I5 can get, and I'm sure you have your pet peeves, too.) Lots of truck driver jobs will disappear, in time, but there will be new sorts of jobs, too, especially for people who keep track of the trucks en route and who maintain them--causing a sea change in workers' comp.

And those are just the changes that pop to mind. Many more that I don't currently envision will come to pass, too, many with big implications for insurers. We're just getting started.

I realize that many people, sometimes including me, have written enthusiastically for years about the prospects for driverless cars and that they haven't exactly rewritten the rules of transportation just yet. The technology actually works incredibly well -- in a world without unpredictable human drivers and pedestrians. But AVs have to fit into that world, and, when they mess up, as they sometimes do, their errors draw a laser focus that crashes involving two human drivers almost never do. 

By some measures, AVs are already safer than most human drivers are, but they are going to have to win the public over as robotaxi services in a few cities, then more cities, then just about every city before we can even start talking about having private ownership of AVs, which is the true mass market and which is where all the rules of transportation and insurance will change.

Driverless big rigs will likely face even more regulatory hurdles than driverless cars do. As an article in the Washington Post notes, there is currently a patchwork of auto regulation state-by-state and little regulation of autonomous vehicles at the federal level. (That will sound familiar to insurers.) AV makers have freedom to operate in any state that doesn't expressly forbid them -- but interstate trucking is heavily regulated, and 35,000-pound tractor-trailers are much more potentially destructive than cars, so states, and likely the federal government, will surely get involved as the AV truck trend gains momentum. 

You never know what regulators will decide. And any mistakes, especially early on, could lead to restrictions on an AV company or even the cancellation of their right to operate; as happened to Cruise in San Francisco when one of its robotaxis dragged a pedestrian 20 feet after a freak accident.  

And yet... trucks operating on highways and freeways are in many ways a better fit than passenger vehicles operating in cities. Cities pose all sorts of complications -- traffic lights, pedestrians (including jaywalkers), double parking, construction and more -- while trucks on open roads face far, far fewer complications. 

There is also a much stronger economic argument for driverless trucks. Driverless robotaxis will eventually become somewhat less expensive than Ubers and Lyfts because a human driver won't need to be paid, but driverless trucks aren't just saving on the driver. They are keeping the trucks moving at all hours, so you won't have an expensive tractor-trailer and all that freight sitting by the side of the road for 13 hours a day on long hauls. 

In addition, the two companies that say they'll roll out fully autonomous big rigs this year are taking a measured approach, unlike some of the autonomous car companies. Elon Musk has repeatedly promised since 2016 that full autonomy for Teslas was just around the corner, only to repeatedly backtrack and lose credibility. Cruise, in what now looks like a show of hubris, decided to tackle San Francisco as its first market for robotaxis. The General Motors unit figured that showing it could handle the steep hills and winding, narrow streets would convince the world that the company's AVs could operate anywhere. Uber made the same sort of bet when it centered its AV operations in Pittsburgh, before ending the program in 2020. 

By contrast, Aurora Innovation and Kodiak Robotics have been testing their truck AI in Texas, where the freeways are in better shape than the roads in old cities and the weather is better. Aurora has been making about 100 deliveries a week for FedEx, Uber Freight and others. According to the Washington Post article, "By the end of this year, Aurora says it plans to have about 20 fully autonomous trucks working the 240-mile stretch between Dallas and Houston. Eventually, it plans to operate thousands of trucks all across America. Kodiak Robotics... similarly plans to launch a fleet of trucks by the end of the year in Texas." 

Since the early days of AVs, many pundits have said autonomous trucks could be the first big use of the technology, ahead of cars. The pundits have sometimes waffled as the technology has taken its twists and turns, and it's still not clear where the commercial breakout will first occur. While Cruise has lost its right to operate robotaxis in San Francisco for now, GM is still investing heavily in the business, so it will be back. Meanwhile, Waymo continues to operate fully autonomous vehicles in San Francisco, recently began service in Los Angeles and says it will open for business in Austin this year. Waymo also keeps expanding in Phoenix, its oldest area of service: The autonomous ride-hailing service now covers 225 square miles, and in January Waymo began taking passengers on highways for the first time. In other words, progress on self-driving cars has hardly stopped, even though the bad press about Cruise might suggest otherwise.

Whenever autonomous trucking kicks in, expect the implications to be profound. The Washington Post quotes Steve Viscelli, a sociologist at the University of Pennsylvania who studies the trucking industry, as saying autonomous trucking could 'change the geography of our economy in the way that railroads and shipping did.'”

Cheers,

Paul

Top articles surrounding our March FOCUS on Underwriting. Plus: an insightful interview with Duck Creek's Jess Keeney.

FROM THE EDITOR  We seem to be in a moment where we're taking a deep breath on generative AI and large language models like ChatGPT. The enthusiasm over the past year and a half has been extraordinary, and we can all see how AI will at least improve productivity and perhaps make more revolutionary changes in insurance. But we haven't seen any killer apps like the electronic spreadsheet that sold so many Apple II's in the late 1970s and 1980s and launched the personal computer revolution. And we're waiting. We're not in what Gartner would call the Trough of Disillusionment on their Technology Hype Cycle. But we are in a Trough of Where Do We Go From Here? When in doubt on breakthrough ideas, I often turn to my friend John Sviokla, whom I've had the pleasure of working with for more than a quarter-century. After years of teaching at Harvard Business School, John in the late 1990s became vice chairman of Diamond Management and Technology Consultants, where I was a partner. He stayed on after PwC bought Diamond in 2010 and was a leader in the strategy and innovation practice, before becoming the marketing lead in the U.S. A year ago, he and some colleagues formed GAI Insights to help companies think about the implications of generative AI and about how to begin implementing it, with insurance as a major focus. They have leaned in to their connections with Harvard, publishing two recent articles in Harvard Business Review; John and one of his colleagues have also become executive fellows at the business school. At Diamond, we joked that we needed to assign a young associate to just follow John around with a grocery sack and collect all the ideas he spun off every day and didn't have the time to follow through on. So, while John and his colleagues are delving into many ideas and developing them deeply, I thought I'd get out a shopping bag and ask John to toss out some ideas about the future of generative AI.  He began with some impressive examples I haven't seen elsewhere of how generative AI is already driving major improvements, not just being tested as part of a proof of concept. For instance, he said, for the past year Jerry has been "using generative AI to handle customer service via chatbots and text. They went from just over 50% of people getting responded to within a day to 100%. Most now get a response within 30 seconds. They've gone from 100% of issues going to a human being to 89% going to the robot—and the 11% that remain with humans get there faster for obvious reasons.... They can grow this business without having to grow the customer service function. ROI is about $4 million a year." He said we've just started to tap the potential of generative AI, partly because of its ability to handle unstructured data. It doesn't just look at data that's perfectly formatted and keyed into traditional insurance data bases; it can draw information from paper or PDFs, from video, from audio, from social media and so on and summarize it in ways that fit into the workflows of agents, underwriters and claims professionals. "We’ve had 120 years of automation of the physical world, 67 years of automation on structured data and 18 months of automation of unstructured data," John said. "There’s lots of opportunity."  He also laid out some thoughts about how to get started, based on where he sees the low-hanging fruit and on a learning model based on the Quality movement. I suspect you'll find the conversation as interesting as I did. Because I subscribe to Amara's law, which says technology breakthroughs are overstated in the short term and understated in the long term, I'll also point you to some reports that suggest we're getting ahead of ourselves.  A survey of Fortune 1000 IT leaders found that 98% had paused generative AI projects because of concerns about security. Another report finds that "adoption is being slowed, primarily by fears around the security and privacy of the technologies (40%), regulation issues (37%) and the skills gap to implement the technologies in house (36%)." But even those reports suggest the problems are temporary and the long-term opportunities are bountiful. So I'm still Team Sviokla at this point. Cheers, Paul   John Sviokla stresses the importance of adapting work processes for AI integration, noting, "You have to redesign work to account for the fact that employees have a new conversation partner, one made of silicon." This change requires understanding and formalizing new interaction patterns for effective collaboration, resulting in a "turbocharged conversation." Read the Full Interview "A lot of times, underwriters are working on descriptions that are pretty thin, that are historical. Well, this is about enriching those descriptions—understanding the semantics, understanding the functional interaction, being able to probe and assess all the dimensions of a risk. You're really expanding the bandwidth of underwriting radically. " — John Sviokla Read the Full Interview READ MORE How AI Can Humanize Insurance AI lets insurers make connections and draw insights from data that otherwise may not have been available, at least not at speed or scale. Read More 'Modernizing Commercial Auto Insurance Advanced data analytics and machine learning can bring insurers back into profitability after years of posting losses. Read More A Reality Check for Generative AI It may require a 50X increase in processing workloads, which means P&C insurers need a modern, cloud-based infrastructure. Read More Can AI Solve Underlying Data Problems? Forward-thinking insurance agencies are ready to put AI to work, but for many, the data just isn’t up to the challenge. Read More Breaking the 'POC Purgatory' Barrier in AI Emerging benchmarks will let firms make informed decisions on whether to scale, streamlining the AI implementation process. Read More How Cameras Transform Workers’ Comp Cameras and AI outperform human observation, which has limitations due to lack of time and inability to objectively measure improvement. Read More FEATURED THOUGHT LEADERS Jess Keeney Patrick Foy Laura Drabik Jennifer Carroll Dustin Ping Jacob Geyer

Dr. John Sviokla is co-founder of GAI Insights. He previously was a strategic adviser at Manifold and former senior partner and chief marketing officer of PWC. He has almost 30 years of experience researching, writing and speaking about digital transformation — making it a reality in companies large and small. He has more than 100 publications in many journals, including Sloan Management Review, WSJ and the Financial Times.

Paul Carroll

To jump right in: You and your colleagues are living on the cutting edge of generative AI, so what are some of the latest examples you’ve seen of powerful uses?

John Sviokla

There’s amazing work being done at Blue Cross/Blue Shield of Michigan. About seven years ago, they began moving beyond functional excellence to build a platform that set them up to offer all sorts of new services to keep customers healthy. 

Ten years ago, they probably had 1% or 2% government business. Now it’s more like a third, and something like 30% of the premium dollars for Medicare Advantage are going to preventative care. That can even mean things like helping pay a light bill or getting someone a pedicure, because it turns out that when they give away pedicures they discover necrosis [dead flesh] far earlier, before it can lead to all sorts of complications. They save hundreds of thousands, if not millions, of dollars. 

When AI came along, they were primed to take advantage. They now have three generative AI products. One gives customers an idea of what benefits are available to them and allows for member self-service. This is super important, given how many benefits there are. One improves members’ security profile, which is needed because of increased access. The third provides an ability to review and manage contracts, which are putting pressure on the insurer because of all the new supply of services and providers, both medical and nonmedical, across the platform. 

They’re not only offering those products themselves but are making them available to the other Blues.

Jerry is going wild with customer service. We’ve written up a case study. They have about 5 million customers and help with insurance selection, refinancing and other issues related to cars. Since April of last year, they’ve been using generative AI to handle customer service via chatbots and text. They went from just over 50% of people getting responded to within a day to 100%. Most now get a response within 30 seconds. They've gone from 100% of issues going to a human being to 89% going to the robot—and the 11% that remain with humans get there faster for obvious reasons. 

They're way more scalable. They can grow this business without having to grow the customer service function. ROI is about $4 million a year.

Drug companies say generative AI has decreased the amount of time from discovery through clinical trials by 60% to 80%. J.P. Morgan claims over 300 productive capabilities. Studies from BCG, Harvard, MIT and so forth find increases in task productivity of anywhere from 10% to 15% at the low end to as much as 50% to 60% on software. 

Paul Carroll

You and two colleagues recently published a piece in Harvard Business Review about how companies aren’t designed for generative AI. What do they need to do so they can accommodate these capabilities?

John Sviokla

The big thing is that you have to redesign work to account for the fact that employees have a new conversation partner: one made of silicon. 

To design for that new dialogue, you have to understand the task at hand, work to discover the new interaction patterns between employees and AI and then codify those patterns and spread the frame. But this isn’t the same as a linear decomposition of tasks, which is how most systems development is done. This is much more about shared discovery, because as the individual is doing the work their job is changing. 

They’re conversing with a different sort of conversation partner, one with a hive intelligence and encyclopedic knowledge. And the conversation is codified into the knowledge base. 

We've got a turbocharged conversation.

Paul Carroll

How do you recommend people get started?

John Sviokla

I think you get started with low-risk parts of the organization. That’s certain kinds of customer service and certain kinds of inquiry, such as people looking for a job or employees with questions on benefits or policies. 

Then you think of this like the Quality process. You want a certain number of white belts, green belts and black belts. White belts, very much like in Quality, have some basic understanding, such as what an LLM [large language model] is and how prompting works. They need to know how to train a model. A green belt knows how to teach a white belt and has managed at least one project to completion. Black belts get more into the technology. The white belts are basically smart users, green belts are kind of in the middle and black belts are closer to people who actually start to build stuff. 

Paul Carroll

You’re always working on the next idea. Where do you think you’ll go next?

John Sviokla

We need to get past sequential processes and understand how to redesign team-based work, so I’m thinking about collective cognition and how to take advantage of our new partner, the machine. We’ve had 120 years of automation of the physical world, 67 years of automation on structured data and 18 months of automation of unstructured data. There’s lots of opportunity. 

Paul Carroll

I get excited about all the unstructured data that’s becoming available. I see companies using AI to make claims or underwriting more efficient, but getting access to all the unstructured data out there will take us to new levels of understanding.

John Sviokla

A lot of times, underwriters are working on descriptions that are pretty thin, that are historical. Well, this is about enriching those descriptions—understanding the semantics, understanding the functional interaction, being able to probe and assess all the dimensions of a risk. You're really expanding the bandwidth of underwriting radically. 

Paul Carroll

Thanks, John. It’s always great talking with you. 

Process: a series of actions or steps taken to achieve a particular end.

From time spent in the military, I will always remember the importance of process. Without process, there is no consistency to approach, no guardrails when times are tough and no foundation to measure against and improve upon. Having moved to the civilian workforce – from the financial industry to the shifting sands of cyber security within the insurance space – I see that that same principle applies. The ability to formalize processes is critical.

In 2024, the cyber situation is dire. According to a report from ESG and the ISSA, 70% of cybersecurity professionals say their organization is affected by the shortage of cybersecurity skills . Meanwhile, the global cyber insurance market tripled in volume in the five years that ended in 2022, and the Insurance Information Institute reports that global direct written premiums for cyber insurance are projected to escalate to $23 billion by 2025.

A skills gap. A surge in premiums. A growing demand. Something has got to give. The world of cyber resilience (which includes cyber insurance) continues to reinvent itself as it changes to mitigate new threats and bad actors. The industry must maintain pace with this evolving landscape. Process is at the heart of this equation.

See also: Top Global Business Risks in 2024

Refining policies and grappling with pricing in this industry’s infancy

It wasn’t too many years ago where the relatively new market of cyber security was thrust into the spotlight following an explosion of digital transformation, at a state and national level. 

The first (and very limited) cyber insurance policies were written in the late nineties and early aughts. Cyber insurance policies were historically centered on security liability, data loss and unauthorized access. 

The scope of cyber continues to broaden as digital transformation and connected technologies overhaul every facet of our lives, from museums opening online collections and archives, to restaurants taking a digital, omnichannel tack with orders, and every business in between. You’ll be hard-pressed to find any industry that doesn’t have digital tentacles attached, ranging from bring your own device (BYOD), to remote devices, to digital subscriptions and more. All this to say that what makes up policies and how they are priced are constantly moving goalposts. 

As cyber insurance matures, and businesses and carriers better understand pricing and risk, this equation will (we hope) balance out. However, to offer more personalized policies and align coverage with unique business demands, the industry will have to shape up quickly. Carriers should look at their underwriting, policy and claims processes to see where efficiencies can be found, where new data can be leveraged and where creativity can support more competitive products. This isn’t the silver bullet to solve all of the industry’s growing pains, but it’s a welcome start.

See also: Top 10 Challenges for Data Security

Lowering demand through education

It’s also important to flip the script from carriers to insureds. This is a team game, with cyber insurance a borderline non-negotiable for any business, agency or organization. The demand is rightfully high. This shouldn’t change. However, there are steps that security leaders can take within their organizations to help the industry at large. Putting the right processes in place can help the industry – as a whole – grapple with this challenge and potentially lower premiums.

Business continuity and disaster recovery strategies: Too often, organizations have these documents or strategies in place, but they are seldom representative of an actual incident that might take place – let alone practiced in such a way. This needs to change. Organizations must show a holistic approach to cyber resilience, from continuing training to technical support to flexible policies that support recovery and damage mitigation. This extends beyond your organization’s four walls and into your partner network, as well.

Balancing domain knowledge with technology expertise: Security leaders will be tasked with showing that they understand not only the security landscape but the business landscape they are supporting. Simply doing security for what is believed to be security’s sake is a path to trouble. The better security leaders can articulate how their processes, programs and protocols map to the specific business challenges at hand, the better their chances at finding and paying for the right coverage. Make sure your organization – and your enterprise ecosystem – is evaluated on having a blend of expertise.

Better measurement of what success looks like: Part of the industry’s maturation is understanding how to measure or demonstrate value to help paint the full security picture. This information can also be used to better evaluate risk and to find, build and price a policy. Security leaders are taking a more prominent seat at the table. Just as a sales or operations or technology leader shows value through key performance indicators (KPIs) and performance measurement, so should security leaders. A good place to start is stacking up against widely accepted security frameworks and training programs. 

Shifting the conversation from cyber security to cyber resiliency

This has been slowly happening, but cyber resiliency will be the way forward – underpinned by the right processes. Resiliency is about more than that one moment when a problem occurs. It’s what is done before, during and (arguably most importantly) after a threat or failure. 

The cyber insurance, mitigation and resilience conversation is here to stay. And it is far from an easy solve. We must reinforce processes – specifically, collaborative processes – in which all parties involved support the common goal of lower premiums, more accurate policies and resilience through education. 

These are the building blocks to a manageable future.

Artificial intelligence (AI) has transcended science fiction to become a fundamental part of our daily lives. In fields like medicine, we've witnessed AI learning in ways that mimic human cognition, particularly in processing immense datasets at remarkable speeds.

While AI algorithms often surpass humans in data processing capabilities, they lack essential human qualities such as empathy and creativity, which are integral to nuanced decision-making in fields like healthcare. The ultimate aim of AI is to replicate human behavior and perform tasks traditionally executed by humans, but there are significant questions about its feasibility and ethical implications, especially in domains where responsible AI practices are crucial.

A report published by McKinsey in early 2023 suggests AI could automate up to 30% of work hours for U.S. employees by 2030, with a more moderate impact expected in healthcare. The report indicates that allied health professionals may see a 4% to 20% increase in automated tasks, while healthcare professionals could expect up to 18% of their work to be automated with AI by 2030. Rather than view AI as a replacement for human expertise, we should see it as a complement—a synergy between humans and computers that leverages the strengths of both. 

Many, myself included, believe that the optimal interaction between humans and AI in healthcare involves a blend of human expertise and AI augmentation. This balance can automate routine tasks that often burden healthcare professionals, such as electronic medical record (EMR) documentation, administrative reporting and even triaging radiology scans. Additionally, AI can assist in analyzing large datasets, providing valuable insights for physician oversight and decision-making.

See also: Streamlining Medical Record Reviews Via AI

AI stands as a transformative force, offering significant advancements in the operation of medical devices and diagnostic capabilities. Deep learning algorithms, for instance, have demonstrated the ability to interpret CT scans at a pace far exceeding human capacity. 

Despite its immense potential, AI in medicine encounters several hurdles that warrant careful consideration. Privacy concerns loom large, as the use of patient data for AI algorithms raises questions regarding data security and patient confidentiality. Moreover, ethical biases embedded within AI algorithms pose a significant challenge, as they have the potential to perpetuate or exacerbate existing disparities in healthcare delivery. 

One of the most pressing issues surrounding AI in healthcare is the lack of comprehensive regulatory oversight. Unlike medical devices or pharmaceuticals, AI software is dynamic and continuously evolving, making it challenging for regulatory bodies such as the FDA to monitor and oversee effectively. As AI technology advances rapidly, regulatory frameworks struggle to keep pace, resulting in a regulatory landscape that is fragmented and often inadequate. 

In response to these challenges, proposals for public-private assurance labs have emerged. These labs would serve as independent entities tasked with assessing the safety, efficacy and ethical implications of AI applications in healthcare.

My journey into the realm of AI was marked by collaboration and a deep dive into the complexities of medical diagnostics. Teaming up with experts from Harvard Medical School, we embarked on an ambitious project to integrate AI with surface electromyography (EMG) readings, aiming to enhance diagnostic accuracy and efficiency. Initially, the allure of AI lay in its potential to streamline the interpretation of surface EMG data, a task traditionally requiring specialized expertise. However, as our endeavor progressed, we encountered the intricate nature of AI and its application in medical diagnostics. 

Surface EMG interaction involves a multitude of variables and considerations,  which needed to be broken down into discrete steps. From analyzing muscle activity to interpreting data points, each step in the process presented unique challenges that AI alone struggled to overcome. It was the human guidance for the AI process that was necessary for the better long-term outcomes. 

The complexity inherent in muscle activity analysis necessitated a comprehensive understanding of various factors, including muscle groups evaluated, movement expectations and spatial-temporal and functional  integration. It became evident that AI, alone, while powerful, lacked the nuanced insight and contextual understanding inherent in human decision-making. 

As we delved deeper into the nuances of AI integration, we recognized the critical role of human guidance in the process. Unlike conventional algorithms, which may operate within predefined parameters, AI in the medical domain demands monitoring and refinement. Factors such as medical history, demographics and individual characteristics must be meticulously accounted for to ensure accurate diagnosis and treatment recommendations. Achieving a seamless integration of AI into medical practice requires not only technological prowess but also human expertise to navigate the intricacies of patient care. 

Our journey with AI underscored the importance of recognizing its limitations and the indispensable role of human involvement in shaping its evolution. This early exposure to AI in the context of musculoskeletal (MSK) conditions underscored the intricate nature of developing AI solutions in medicine. While AI holds immense promise in revolutionizing medical diagnostics, its efficacy ultimately hinges on the quality of data inputted and the oversight provided by human experts. 

Our AI integration road map is grounded in a comprehensive understanding of the complexities inherent in EMG integration and medical indications. Collaborating with experts in these fields, we prioritize a phased approach to algorithm development, recognizing the need for iterative refinement. 

See also: Data Science Is Transforming Public Health

In my experience, the choice between deploying AI or relying on human expertise necessitates a thorough consideration of unintended consequences. While AI presents vast potential, its reliance on training data introduces the risk of bias, a factor that often goes unnoticed in decision-making. Moreover, AI exhibits a slower adaptability to unnecessary changes, lacking the imaginative and innovative capacities intrinsic to human cognition. Humans, on the other hand, possess a distinctive aptitude for exercising discernment, multitasking proficiently and comprehending information in nuanced ways beyond the capabilities of machines. 

The integration of AI into healthcare holds immense potential to improve efficiency and outcomes. However, it must be approached with caution and a keen awareness of the ethical considerations involved. By embracing a collaborative approach that combines human expertise with AI augmentation, we can harness the full potential of technology while prioritizing patient care and safety.

The rate at which data science techniques are developing and being adopted is increasing faster than insurers are able to develop their own understanding of the risk governance and ethics needed.

To make matters more challenging, two distinct groups operate within most insurers on the front line of data science, often in conflict rather than in harmony: data science teams using cutting-edge techniques without the necessary understanding of their organization’s risk frameworks, and insurance leaders who have limited experience with the latest advanced analytics. This internal disconnect leaves insurers and individuals that work for them exposed to risk.

Finding the right balance between governance and control, while still advancing the adoption of data science and the value that it creates, has become the magic middle ground upon which insurers have set their sights. 

See also: Why Becoming Data-Driven Is Crucial

Bias

As increasingly complex models are used, a key risk for insurers to consider is bias -- an issue so far  not fully understood and appreciated by many firms. When individuals or groups are differentiated from others based on particular characteristics, insurers need to understand why. Is the bias due to the data collected not representing the entire population? Is it caused by potentially flawed human decision-making that is represented in the data collected? Or was the bias introduced due to the artificial intelligence (AI) and machine learning models trained on the data? Is the inherent model form being used responsible for reinforcing the existing bias or even creating new biases?

The ability to detect hidden biases is essential to enabling appropriate strategies to measure, monitor and manage bias. Instead of thinking about bias at every stage of the model building process -- when an insurer first explores their data, when they build a model and when model outputs are used in a business decision – data scientists too often consider the risks as an afterthought. 

Choosing the right algorithm that will help an insurer find the optimum balance among interpretability, transparency and predictive power is another essential capability. There are a number of custom algorithms being developed in the market. For example, layered gradient boosting machines (LGBM) capture the same predictive accuracy of a GBM, while providing a much greater level of transparency and interpretability.

Open source risk

In recent years, open source adoption has seen unprecedented growth. While open source allows incredible flexibility and innovation, it also exposes an insurer to more risk, particularly relating to governance and security. Besides the potential for malicious code hiding in open source packages, key person dependency is another risk created by having either just one individual or a small team responsible for building and maintaining code. 

Large language models (LLMs), such as ChatGPT, are examples of technology evolving and being adopted in a hurry. However, the governance risk and control frameworks have not kept pace, creating significant risks relating to data privacy and intellectual property. 

Through the use of LLMs, an insurer could potentially lose sensitive and proprietary data. There is potential to have no or limited control over how the data is used, including being used by competitors later. 

Another risk concerns hallucinations, which refer to the tendency of LLMs to produce text that appears to be correct but is actually false. This could be driven by bad prompts or simply due to an underlying weakness in the model, delivering results that are wrong but are presented with a lot of certainty. Reputational risk for an insurer is high if the data or model is used improperly.

See also: Data Science Is Transforming Public Health

Taking control

At the end of the day, the stability of the open source code is in the insurers’ own hands. They alone are responsible for making sure they meet their business needs. Therefore, it is important that an insurer clearly delineates roles and responsibilities to avoid confusion. Defining who is making which decision ensures that better accountability, visibility and opportunity to challenge decisions are in place at every level. 

Open source offers real potential to contribute to a more efficient and innovative insurance market. However, insurers must first address two critical decisions: what they should use open source for to gain an advantage; and then how best to integrate open source in such a way that good governance and control are in place, creating an optimal balance. 

Data science is spreading quickly. If insurers want to compete in this new AI-driven world, they not only need to simply adopt data science but also do it in the right way. This means a gradual evolution of governance to ensure the right oversight, alignment to internal values and regulatory compliance are achieved, combined with an evolving risk management framework to anticipate and mitigate future risks.

The first quarter of the year just ended, and so far 2024 feels a bit different than 2023. U.S. economic indicators are fluctuating month to month, and policymakers feel we are not out of the recessionary woods yet. The insurance industry has struggled to stay profitable under destructive weather events and ballooning costs. Insurers are striving to adapt to the new normal of continuous crisis, continuing digitization and evolving customer expectations. 

1. Developing Risk Resilience

In the middle of winter, firefighters raced to contain the Smokehouse Creek Fire, the second-largest wildfire in Texas history, raging in the panhandle. As we witness this early challenge posed by natural disasters, it prompts us to contemplate what lies ahead as we approach other extreme weather seasons, such as hurricane and wildfire seasons. After years of record natural disasters ushering in the era of global warming, the insurance industry is now bracing for further calamities. These unanticipated risks are not the only areas insurers need to watch out for. There is a new breed of non-natural disasters, ranging from cybersecurity to machinery malfunctions (such as Boeing's safety issues), that have prompted insurance companies to look for better ways of managing these exposures and enhancing risk resilience. 

To fortify risk management, insurers should combine traditional practices with advanced technology tools. These tools enable insurers to monitor and stay abreast of evolving conditions while efficiently gathering data to identify risks and vulnerabilities. By integrating technology into their risk management strategies, insurers can streamline processes, improve data analysis and address emerging threats, ultimately enhancing their ability to mitigate risks and protect their business and policyholders.

See also: 20 Issues to Watch in 2024

2. Insurance GPT: AI’s Impact

If 2023 proved anything, it was that large language models (LLM), aka generative AI, are not going anywhere. Insurance companies must understand the numerous use cases AI can have on the industry, specifically on underwriting, insurance product sales and customer service. As in most sectors, gen AI tools are great for freeing up various functions for higher-value tasks like risk mitigation, by automating numerous manual tasks, from summarization of documents to synthesizing large volumes of data. 

Adoption of the LLMs is improving workflows, boosting productivity, and unearthing efficiencies. It may be a while until we understand the depth of impact of LLMs on our industry, but insurers will need to integrate them into our daily practices sooner rather than later to stay competitive. Gen AI is in its infancy, where everyone is essentially a beginner, but experts are emerging. This presents a remarkable opportunity for organizations to carve out a niche and become leaders. For instance, Marsh McLennan has become an early adopter and launched a generative AI tool last year to help its 85,000 employees worldwide to streamline their work processes. Moreover, in an industry facing disruption, busy trying to recalibrate its role in society, LLMs could contribute to transforming the customer experience (CX) game.

3. Rate Increases

Premiums have already been increasing for four years, which makes elevating CX much more critical. Both the $32 billion in P&C net underwriting losses in January-September 2023 and P&C net combined ratio of 103.9% are decidedly unpleasant numbers. Fitch Ratings regards 2024’s outlook as “neutral,” and “the market combined ratio is projected at slightly over 100%.” While insurance companies are reinventing themselves with a laser focus on closer customer engagement, they know they need to increase premiums. Insurers are operating in a Catch-22 of sorts. They are feeling the pressure to reduce costs and raise rates in personal and commercial lines within insurance policies – but it’s hard to justify without adding value and becoming more transparent. Leveraging generative AI and other analytics technologies, combined with partnering with insurtechs, may be key to finding operational efficiencies.

4. Managing Social Media

While AI chatbots and other productivity tools are infiltrating every aspect of our lives, social media has not exactly faded into obscurity. In fact, consumers are turning to social media channels and online sources as an information source to dig for details on matters that affect their daily lives. With social media channels remaining as an important touchpoint, insurers still need to provide useful content across all communications channels to guide customers along their buying journey. This can prove problematic, hampering the underwriting process when consumers turn to social media to understand insurance policy pricing. This is a delicate line to walk as insurance companies cannot neglect their social media presence and must ensure their processes are completed properly. Conversely, insurance companies can still glean actionable data from consumers interactions with social media. 

See also: Insurance in 2030: What Does the Future Hold?

Steering the Year Ahead

Dealing with the unexpected is a routine challenge in the insurance sector, and need to do so is likely to escalate in 2024. To navigate successfully, insurers must prioritize adaptability and adjust their strategies to match the dynamic landscape of the industry. Maintaining an aggressive approach to risk management and embracing innovation and technology will be essential for staying ahead of emerging challenges and seizing new opportunities.

This week, I'll quickly hit three themes, including an audacious new form of cyber attack and General Motors' decision to back down on sharing data with brokers from GM vehicles, in the face of a recent uproar. But I'll start with a step for homeowners insurance toward a "Predict & Prevent" business model and away from the traditional repair-and-replace approach.

Realtor.com said it will provide projections on climate risk for each listing. Heat, wind and air quality scores will predict what an area will experience, going out 30 years, supplementing existing data on fire and flooding, so home buyers can know more about what perils -- and high insurance premiums -- they may face as the years go by, BEFORE they make the purchase. 

Economists at Realtor.com say some $22 trillion of U.S. residential properties are at risk of ‘“severe or extreme damage” from flooding, high winds, wildfires, heat or poor air quality. And First Street Foundation, which provides the data that is the basis for Realtor.com's analysis, says nearly 36 million homes -- a quarter of all U.S. real estate — face rising insurance costs and reduced coverage options due to mounting climate risks, First Street Foundation, a nonprofit that studies climate risks, says 18% of U.S. homes will be at risk of damage from hurricane-strength winds over the next three decades.

The projections will be far from perfect, especially in the early days, but the hope of the Predict & Prevent model is that we can keep adding intelligence to the system and help customers make decisions that let them avoid putting themselves at risk, long before a flood or fire or windstorm can damage or ruin their property. (If you're interested in learning more about what Realtor.com is doing, or about Predict & Prevent in general, I encourage you to check out The Institutes' Predict & Prevent newsletter. You can also sign up for regular podcasts on Predict & Prevent.)

Now on to the brazen new cyber attacks....

The Wall Street Journal reports that cyber criminals have gone far beyond stealing data and are now stealing data that they use to then steal physical goods -- they hack digital exchanges that facilitate shipping via truck and divert loads of products that they hold for ransom. The WSJ says reports of freight fraud more than quadrupled in 2023 from the year before, and fraud-related losses were estimated to be at least $500 million last year.

Criminals hack into an exchange that those shipping products use to coordinate with trucking companies and steal the unique identifier of a trucking company. They use that identity to submit a lowball bid and win the right to ship a truckload or more. The criminals then subcontract with an actual trucking firm that, not knowing it's participating in a crime, picks up the load or loads and delivers the goods to the hackers. The hackers contact the company that owns the goods and demand a ransom. Often, the goods are perishable, such as yogurt, and some store chain is angry about not getting its scheduled delivery, so a ransom is paid before authorities even have a chance to start unraveling the crime.

Everybody seems to be pointing fingers at everybody else. Those in the shipping industry say authorities need to do more to stamp out the fraud. Authorities, who say they don't have sufficient resources, say shippers, brokers and truckers could do a lot to police the problem themselves by watching out for bad actors and coordinating to kick them out of the system.   

Trucking capacity was so scarce as the economy heated up following the worst of the pandemic that many shippers and brokers didn't pay enough attention to making sure that a trucker was legitimate, and fraud may abate as everyone -- including insurers -- pays more attention. Here's hoping. But in the meantime, freight fraud has become a major problem. 

As for General Motors: Following loads of negative publicity (including from yours truly) about customers complaining that they were being tracked without their approval, leading to higher insurance premiums, the company said it would stop sharing data on drivers with data brokers such as LexisNexis and Verisk. That is a smart PR move, but I don't think the issue will go away. GM still faces a lawsuit that the plaintiff hopes will be certified as a class action, and the whole issue of tracking drivers, perhaps without their consent, seems to have drawn the attention of the Federal Trade Commission and of Congress.  

Stay tuned.

Cheers,

Paul

As the first quarter of 2024 winds down, the insurance industry is embracing digitization and recognizing AI’s great promise for improving and simplifying what we do. It’s exciting to contemplate all the potential that technology affords. Yet we can’t lose sight of the fact that technology will never fully replace human connection.

Finding the best ways to use technology is understandably a focus for our industry, but the key to success lies in finding ways for technology, soft skills and empathy to work together. This is best done by investing in technologies that automate lower-level tasks and make employees’ jobs easier, which helps to clear the way for more high-quality interactions with clients and customers. 

Equally important is finding ways technology can be used to cultivate a culture of empathy internally. A workplace empowered by empathy and efficiency allows employees to build strong relationships with both clients and policyholders. By combining human connection with technology-enabled solutions, claims professionals will have more time for fully engaged customer experiences. Think of this as human intelligence supported by artificial intelligence, not as artificial intelligence replacing human intelligence.

Opportunities to improve claims process efficiency with technology should not come at the expense of the human element. 

Insurance is a product that nobody ever wants to use. But claims are inevitable, and when they do occur it’s important that policyholders feel that their insurer is there for them on a human level. If your home or your car is damaged, or if your business operations are on the line, you may not want a robot or an automated text message telling you that everything will be okay. Rather, you want a person to reassure you and give you confidence that things will be taken care of. For a policyholder or claimant who’s just gone through potentially one of the most terrible days of their life and wants to be shown empathy, reassurance is part of the product – and that reassurance needs to come from a person, not a bot.

In contrast to claimants looking for the human touch, technology enablement is perfect for the policyholder primarily looking for simplified self-service solutions or higher degrees of digital engagement in their claim experience. We must be careful, however, not to miss the mark by assuming that we know what someone needs or wants. A claim is a personal experience for everyone, and our role as claims professionals is to meet each customer where they are. With technology, we are better equipped to tailor the approach down to the individual preferences or needs of the customer and their circumstances. The integration of self-service interfaces to the claims process gives each insured the ability to define their own journey – and their claims team the ability to deliver the experience each customer needs and expects. 

See also: Adding Humanity to Life Insurance

Beyond process efficiency, empathetic claim management can also help avoid large jury awards and nuclear verdicts.

Just as the impact of biases can work against businesses, a good claims experience can work in an organization’s favor. With social inflation and nuclear verdicts currently creating challenges for the insurance industry, it’s worth noting that a little bit of compassion can have a significant impact in combatting the anger and fear driving these trends. Claims organizations can avoid litigation and achieve quicker, less costly settlements when they take an advocacy-based approach that prioritizes care for injured parties. It stands to reason that jurors who see the insurance company as caring and empathetic may be less likely to punish them. 

Critical components of an advocacy-based model include ensuring early outreach, offering empathy and support regardless of fault and going above and beyond to demonstrate care. Each of these “human” behaviors can go a long way toward mitigating confrontational litigation. This is all the more reason to resist the urge to over-automate larger value claims processes and to keep empathetic personal connections.

Empathy can be more profoundly extended to clients if it is first demonstrated to employees.

Not only should we look to balance automation with human support for policyholders, but as employers, we must also strive to provide empathy to our own people. Adjusters’ behaviors can change depending on what’s going on in their personal circumstances, and we as an industry must respond to these shifts by showing empathy to our own people, as we ask them to show empathy to others. 

Creating a workplace empowered by empathy is essential to building a resilient workforce. Using human connection as a lens through which day-to-day business practices are viewed and implementing technology that frees up more time for relationship building are two key ways that organizations can sustain and improve empathy in the workplace. Workers aren’t going to be able to focus on empathy when they’re struggling, so it’s essential to prioritize holistic health and ensure that employees can manage the stress that accompanies helping customers in difficult circumstances.

See also: 4 Key Questions to Ask About Generative AI

The goal for 2024: Make empathy and technology work together to empower employees and drive client satisfaction.

As technology matures, it offers increasing opportunities to streamline workloads, allowing employees to focus on clients' claims with more attention and empathy to deliver personalized customer experiences. 

The newest tools offer even more promise for blurring the lines between technology and empathy. AI algorithms can now identify and categorize emotions by analyzing text, speech and other data sources, and this information can be used to increase empathy even in tech-driven interactions. AI-powered sentiment analysis can help organizations to analyze and understand customer emotions, attitudes and opinions, as well as to monitor social media feeds, customer service interactions and customer feedback surveys to identify issues requiring additional response. These tools can also be leveraged to help companies identify bias in their customer interactions. And new call center tools are available that provide real-time coaching and flag potential issues, allowing the agent to focus on the caller, not the script. These tools could also flag the agent to signs of anger, frustration or confusion with the caller so they can course correct during a call.

When human connection is combined with technology-enabled solutions, claims professionals have more time for fully engaged customer experiences. And when employees are feeling their best, they will have the patience and soft skills necessary to be present and thoughtful with clients and customers. Technology that supports and builds time for human interaction should be the priority for the claims industry and will be what drives the industry’s continued success in the coming year.