Fifteen years ago, I was working for a company with offices in California, New York, and London. I was at a conference in the U.S. My wife was back in London, with our three young children while overseeing the renovations of our 300-year-old house.  

While I was away, the kitchen ceiling collapsed, bringing down with it an old water tank and soaking everything in dirty water. The au pair quit the same day. I hadn’t been in touch for 48 hours. My wife didn’t even know where I was staying. 

So she sent an email to Matthew Grant – expressing some frustration that I hadn’t called and needing sympathy for the troubles at home. An hour or so later, she received a response:

“I’m sorry but I don’t know who you are.”

She saw red. 

How dare I pretend not to know her because I was at work and she was having a hard time? She shot back a response:

“You B@st@rd, how dare you ignore me just because I’m having bad day. Who do you think you are?” 

A few minutes later, she got a reply

“No, I really don’t know who you are. I am not your husband. But I do hope you sort out your troubles.”

She looked at the email address she had used. She had indeed sent her original email to Matthew Grant at Hotmail. Just not to the right Hotmail address. 

We laugh about it now, but we both learned something about communication that day.

See also: Life Insurers' Communication Problem

Don’t be a stranger…

I was reminded of the comment, “I don’t know who you are,” when looking this week at the backlog of all the people who have asked to connect to me on LinkedIn. 

I have a something of a love/hate relationship with LinkedIn. 

We use it a lot at InsTech for sharing information about what we are doing and what our clients and members are up to. It’s a great way to share longer-form articles, event information, and links to our podcast episodes.

But none of us can control what the LinkedIn algorithm does with its content. Despite being selective about whom I open up my network to, I still get too much irrelevant information in my feed. And I get a lot of strangers asking to connect. 

LinkedIn favors people from our connections when it shows us content, so one way to manage the quality of what we see is being thoughtful about whom we agree to connect with. 

I also still believe in the original concept of LinkedIn, that connections are people in your network whom you know and whom you have a connection with in the real world outside of the platform. 

I realize this may be more of a personal pet peeve, but I suspect I am not alone. Some people value the quantity of their connections on LinkedIn in the same way others value the size of their bank accounts. I’m not judging that approach if it works for you. For me, personally, I have a bias toward the quality of my connections.  

I tend to get four types of people asking to connect. You probably do, too:

Type One – People I know. A pleasure to be connected, thank you.  

Type Two – People I don’t know but who send a message explaining why they want to connect. I almost always respond to those and accept. 

Type Three – People from companies that are already clients or who are likely to be buyers or users of our services, or the services of  our clients and members. If you work for an insurer or a large technology company, or are an existing client, or perspective client, we need more people like you in our network. Thank you. 

Type Four – People I don’t know and whose employers I don’t know. These people offer no message or explanation about why they want to contact. If I don’t know you, how do I know how I can I help you, or what I have I done that brought you to me? Sorry, but that will be a “decline.”

See also: 3 Steps for Insurers to Keep the Human Touch

Managing the communication overload

We are overwhelmed with the need to communicate today. Sometimes we spend too much time in “receive” mode and neglect the effort we spend in “transmit” mode. We respond to what is present, not what is important. We can write quickly, but we suffer regret slowly. 

Can I offer some advice that might help make sure you are not the stranger and may reduce the likelihood of a small communication catastrophe?

Lesson One: Double check the email address before you send. 

It’s too easy to send an email to the wrong person. I suspect you have done it more than once. I have. Most email tools use text prediction to complete the email address you start with. There are lots of Matthews out there. Are you sure you have the right one?

Lesson Two: Never say anything in an email you write that you wouldn’t want the person you are writing about to read. 

Assume that any email you write about someone will be sent to that person. Because that will happen one day. It’s Murphy’s Law. If something can go wrong, it will. Harsh words said in haste can’t be taken back, and they are really hard to recover from.

Lesson Three: Who are you? Introduce yourself when requesting a connection on LinkedIn. 

If you want to get someone’s attention on LinkedIn, take the time to say hello and introduce yourself. Tell them what they have done that has caught your attention. Don’t make the first contact be a pitch to sell something. 

Lesson Four:  If you travel and have someone waiting at home, check in every single day. 

A phone call is best. Otherwise a text, WhatsApp, or email at least shows you remember them. It’s not every day something troubling happens, but each day does have something worth sharing, the good and the bad. These are the small parts of our day that may be forgotten or seem not important 24 hours later but which form the broader tapestry of how we experience our lives. That daily contact is part of the vital glue that holds us together. 

I am still married to Mrs Grant. Occasionally, the other Matthew Grant does still get emails. I’m assured the offer to book flights to go on holiday in Bermuda that he received was entirely accidental.

Four suggestions. Take them or leave them. If you have read this far, then this is one communication that has succeeded. That’s good enough for me.

How Digital Trust Strengthens Customer Loyalty

In today’s fast-paced world, insurance companies face the challenge of balancing customer trust with quick service. Technology allows quick decisions and real-time service but also makes customers more impatient and demanding. Insurance carriers must meet these needs while being careful, especially to prevent fraud.

Good processes help maintain customer trust. But trust can also slow down services. This blog looks at whether customer interest and care can work together. 

The Shift from Distrust to Trust Automation

In the past, carriers unfortunately had little options than to operate based on distrust, focusing on controlling damage. There are many rules to manage risks and avoid losses. Now, the industry is moving towards each customer’s experience. Modern carriers want to build personal relationships. They use technology to meet each customer's needs. The market is less about groups and statistics and more about serving one unique person well.

Trust. Automated. A Solution for Modern Challenges

Automating trust processes helps carriers manage lots of interactions smoothly. If you can automate who to trust based on real-time data, carriers can act quickly and focus on high-risk cases, while speeding up low-risk ones. By using advanced analytics and machine learning, the platform spots patterns and odd behaviors. 

Key Parts of Trust Automation

1. Primary Verification: Checking personal details, income, and more to ensure accuracy. This step ensures that the information provided by customers is correct, which builds a foundation of trust.
2. Text Mining and Anomaly Detection: Analyzing large amounts of data to find unusual patterns and potential fraud. This helps in identifying risks early and avoiding potential losses.
3. Social Network and Voice Analysis: Mapping social links and analyzing voice signals to judge trustworthiness. Understanding relationships and how people communicate helps in assessing risk more accurately.

Benefits of Trust Automation

Automating trust brings significant benefits for both customers and carriers. 

Customers value quick and efficient service. With trust automation, carriers can offer expedited claims processing for trustworthy clients. This reduces wait times and improves customer satisfaction. Happy customers are more likely to remain loyal, refer others, and even purchase additional policies.

For carriers, the benefits extend beyond customer satisfaction. Trust automation streamlines operations by handling large volumes of data and interactions consistently. Carriers can standardize verification processes, ensuring that all customers are treated fairly. This consistency builds trust and helps in maintaining a positive reputation.

Detecting and Preventing Fraud

As data grows, systems get more complex. This affects decision-making speed and makes fraud harder to spot. Trust automation helps normalize and standardize many processes. It lets skilled staff check exceptions when needed. Examples include:

• Text Mining: Analyzing large amounts of text for patterns and tendencies. This involves scanning through documents, emails, and other text data to find inconsistencies or red flags.
• Anomaly Detection: Identifying items or events that don’t fit expected patterns. This could be unusual claim amounts or irregular policyholder behavior.
• Social Network Analysis: Mapping social structures to find hidden connections. This technique helps in identifying relationships that may not be apparent but could indicate collusion or fraud.
• Voice Analysis: Using voice analysis to catch signals of someone’s state of mind. Voice patterns can reveal stress or dishonesty, providing additional clues during claims processing.
• Financial Risk Analysis: Charting the financial risk of a company or person using public info. This step helps in assessing the overall risk profile based on financial stability and history.

Conclusion

Fraud prevention and customer trust can work together. By using technology to automate trust, carriers meet modern customer needs while maintaining care. Trust automation boosts customer loyalty and makes operations efficient. It is a valuable tool for the insurance industry. As the insurance world changes, carriers must use new solutions like trust automation to stay competitive. Building trust through technology helps carriers create lasting customer relationships and ensure long-term success. 

In summary, trust automation is not just about preventing fraud; it's about creating a more efficient and customer-centric insurance experience. Technology is reshaping the industry, and those who embrace these changes will lead the way in customer satisfaction and operational excellence.

Sponsored by ITL Partner: FRISS

Paul Carroll

From your perspective, how is the role of agents changing?

Chase Tarkenton

We're seeing a paradigm shift with the onset of AI. Automation has taken such a big leap forward that it’s freeing agents to focus on their clients and provide more thoughtful, more personalized interactions and focus on higher-value activities. The agent can provide a better experience, which is ultimately what the consumer is looking for.

For years, technology has been integral to supporting agents, but a lot of new use cases are coming up every month, where you can trust that the technology will execute and do so at scale and in a compliant fashion. It's really exciting to see.

Paul Carroll

If I'm an agent, how might my day look different now?

Chase Tarkenton

Automation is a big topic, so it can touch lots of different areas of the agent workflow.

With the onset of self-service, more and more customers are coming in already educated on different insurance policies. Now, when an agent is interacting with those clients, some of those conversations have already started, and maybe the agent has been provided a summary of how the customer has been interacting with the organization and what their situation is. By the time the agent speaks with someone, whether it's over the phone, over email, or through digital channels, the agent already has context, and they can pick up the discussion midstream.

At the end of the interactions, the paperwork that the agent has to fill out can be automated. Or a survey can be automatically sent to capture the client’s feedback.

Ultimately, customers can self-serve, and if they don't want to automate, there's that human in the loop. That human is the agent, who can have higher-quality interactions and more interactions on a daily basis than they could previously.

Paul Carroll

I’ve been an advocate for self-service for decades, but in my experience the handoffs to humans from the AI can sometimes be messy. How do you tackle those problems?

Chase Tarkenton

Obviously, the quality of the AI matters. So does the quality of the partner providing the AI to the agents and brokers. Unfortunately, there are a lot of technologies out there today that cannot deliver on some of the promises that are made.

Anyone considering a chatbot system should talk to reference customers and do a pilot to spot where problems might arise and to figure out in advance how to handle any issues.

Paul Carroll

I assume that this sort of technology, like most I've covered over the decades, has a progression to it. What are the sorts of interactions you started using AI with? And where are you now?

Chase Tarkenton

Twenty years or more ago, chat and voice virtual agents, as we call them, just handled basic FAQs [frequently asked questions]. What’s your phone number? How do I find a branch location? The agents were useful, but I wouldn’t say they were high value.

Fast forward from there, with the emergence of natural language understanding and the incorporation of machine learning and AI, many of these technologies can now handle more sophisticated use cases. Customers can fully resolve requests without having to call and speak to a human—though it’s an important design consideration to make sure they can reach a human if they want.

For example, look at first notice of loss. If your car breaks down on the side of the road, you could pick up your phone, call your insurance provider, and fully authenticate yourself through voice automation. They could leverage geo tracking to know where you are on the side of a busy highway and call a tow truck to come get you. They could capture information automatically and complete first notice of loss without any need to escalate or require a call center or an agent to support them.

That's a great use case that we're seeing come up again and again, and insurance organizations are executing on it. Their NPS [net promoter score] stays consistent with the NPS of a human interaction, which is what we want.

Paul Carroll

Where do you think you can get in two or three years?

Chase Tarkenton

Things are moving so fast that it’s hard to say what’s two or three years out in the AI space, but we’re clearly going to see generative capabilities produce some very exciting outcomes.

We're seeing models produced by some of the bigger generative providers that can reason and handle complex interactions. Open AI’s latest model does complex reasoning in a scientific setting. Those are breakthroughs that will allow for more intense use cases within insurance.

While we can all get excited about technology, though, what matters most is: What does the customer want, and are we doing everything we can to help them be successful? A close second is: How do we make the agent as successful as possible in their role? How do we make them efficient? How do we make them effective? How do we help them to scale?

I think using pre-trained models that are then trained on the data of the insurance organization, with generative capabilities blended in, can drive hyper personalization, which is what the industry is asking for.

Paul Carroll

Where do you see the breakdowns at this point? Where are people most likely to say, "Okay, I’m done with the chatbot. I want to speak to a human?"

Chase Tarkenton

In the design of an automated experience, there are a couple of things that are just a hard stop for consumers.

Number one, if the bot doesn't understand what you're saying, we have very low patience and will immediately drop out of those channels and want to speak to a human.

Number two: Sometimes a partial experience is not what consumers want. They're looking for full resolution. If I want to check what my premium is and can only do part of that before I’m escalated to a human, where I have to repeat myself, that's a non-starter, right?

Any time an interaction is being escalated to a human, it’s critical that the context and information already gathered is transferred to that agent in advance, so it's not being repeated.

Paul Carroll

Are there any issues that I should have gotten into that I didn't?

Chase Tarkenton

AI and automation are a huge topic. The only other areas that might be worth getting into are security and compliance.

Paul Carroll

Great question. Glad I thought of it. How do you handle security and compliance?

Chase Tarkenton

The good news is that the tools exist to mitigate hallucinations and to make sure that any AI-based agent will speak in a fully compliant fashion. If an agent or a customer asks the bot a question, it can quickly recognize whether responding would violate compliance.

Those tools are being used today. At Boost.ai, we’ve got hundreds of financial institutions, many in insurance, that are leveraging those tools to get the results they’re looking for: a better customer experience that they know is ironclad, protecting their reputation and their brand from a security standpoint.

Paul Carroll

Thanks, Chase.

Succession Planning for Agencies

Severe Weather Needs Innovative Insurance

About Chase Tarkenton

Chase Tarkenton is the SVP and general manager of boost.ai, North America. He’s focused on partnership growth, helping insurance firms leverage AI technology in personalized customer experiences.

While generative AI has been a phenomenon over the past couple of years, most uses have stayed pretty close to home — gathering information to make underwriters, adjusters, and agents more efficient, producing first drafts of reports or communications to clients, that sort of thing. 

But some recent articles suggest that AI may be getting ready to break out into much more sophisticated uses, including showing up as part of insurance products.

The article that most caught my eye was one in Fortune about how the CEO of Honeywell intends to use AI for competitive advantage. I had just finished another article in Fortune about a long list of impressive things that Honeywell was doing with generative AI at the operational level and was startled that the CEO belittled the effort. He said the work had to be done, because all his competitors were doing it, but said all the gains in efficiency would be competed away. To gain a sustainable advantage, he said, Honeywell needed to be bolder.

The CEO, Vimal Kapur, mentioned three areas where Honeywell is focusing, all of which strike me as being opportunities for the insurance industry, as well.

First, he talked about using AI to address his industry's talent gap. That issue sound familiar to anyone in insurance?

Kapur said, "If historically somebody said, ‘This job requires 12 years or 15 years experience,’ well, maybe you’re going to be able to have someone do it with seven years experience, and a supplementer [based on generative AI].... What’s a Plan B? There is no Plan B. There’s no humans to replace the humans who have left the workforce.”

He's talking mostly about engineers, so a different pool of talent than the insurance industry requires, but the issues strike me as very similar. AI can provide tremendous aid to help newer employees operate at more sophisticated levels.

Second, he talked about the area that intrigues me the most: actually adding AI to products, not just using it in the background. 

The article says: 

"He gives examples of supermarket checkout scanners, of which Honeywell is a major producer. Today, these scanners work well for bar-coded products. But if you get to the counter with an individual piece of fruit or vegetables where the bar-coded sticker hasn’t been applied or has fallen off, then the cashier must manually look up the price, or the customer is forced to go and weigh the product individually on a separate digital scale, often holding up the line. Kapur says that integrating cameras and computer vision directly into Honeywell’s checkout scanners would enable the scanner to recognize the item and charge the customer appropriately, without delaying the process."

Again, not an insurance example, but it illustrates the kind of thing that's possible for insurers, especially as more companies adopt a Predict & Prevent business model.

AI is already getting built into insurance offerings: the computer vision that monitors the road and the driver and offers warnings in real-time that can prevent accidents; the sensors that detect water leaks and alert homeowners before major damage can occur; Whisker Labs' Ting, which plugs into a wall socket and detects electrical anomalies and warns policyholders before a fire can start. 

But I can imagine a host of other opportunities, particularly in risk management. Who wouldn't like to have a smart adviser whispering in their ear that a risk is developing. Generative AI can be that adviser on issues as complex as cyber or as mundane as the approach of a hail storm, some crime trend, the need for a roof inspection, or certain home maintenance. 

You don't just sell a policy. You provide some continuing intelligence based on AI that is constantly learning.

If you do this right — delivering insights that are truly smart, that are useful, and that are delivered in the way that policyholders want to receive them — then you open up the sort of line of communication with customers that insurers have long craved. Insurance companies complain that they only interact infrequently and in cursory ways with most customers, when they make their monthly, semi-annual, or annual payments. Adding an AI channel would change all that. 

Finally, the article says another area where Kapur will use AI to Honeywell's advantage "is in providing engineering solutions to customers. In many cases, Honeywell doesn’t just sell an off-the-shelf product. It sells a system incorporating several of its products. These are usually built to a customer’s specifications, a task that requires a significant amount of time from the 5,000 engineers Honeywell employs for this work. 'We write the solution or spec for the project every time and we do tens of thousands of projects in our business every year,' he says.

"Kapur says Honeywell wants to build a large language model that can streamline this spec writing process, so that what currently takes as long as a month could be completed in just minutes—with engineers checking the output for perhaps a few additional days to guard against the risk of AI 'hallucinations.'” 

What he's describing sure sounds like it could be applied to the writing of complex insurance policies.

Another recent article tees up the possibilities of what's sometimes called agentic AI. which I think is a bit further out than the sorts of things Honeywell is pursuing but which could certainly be a profound advance for generative AI. The basic idea is that the AI wouldn't just have the authority to gather and work with information. The AI would also have the ability and authority to work with apps and have them execute tasks on behalf of the user. 

In other words, you wouldn't just tell the AI to gather tips on how to build a website. You'd tell the AI to build a website, and it would. 

The article's author, Bernard Marr, offers a range of areas where agentic AI could make a big difference. For instance:

"Business Operations: Agentic AI could revolutionize how businesses handle day-to-day operations. These AI agents could autonomously manage supply chains, optimize inventory levels, forecast demand, and even handle complex logistics planning. By processing vast amounts of data and making real-time decisions, they could significantly improve operational efficiency and reduce costs."

Or:

"Healthcare: Agentic AI could revolutionize patient care by serving as round-the-clock health assistants. These AI agents could engage with patients daily, monitoring their mental and physical health, adjusting treatment plans in real-time, and even providing personalized therapy support. By analyzing vast amounts of medical data, they could also predict potential health issues before they become serious, enabling truly proactive healthcare.

He also describes opportunities in software development, cybersecurity, human resources, scientific research, and finance. 

Again, I think agentic AI will take a while to take hold, just because a hallucination could lead to such big problems if the AI has the authority to act on its own. But the premise is certainly provocative and should be watched.

We're still in the very early innings in terms of what generative AI will do.

Cheers,

Paul

Let’s consider the $300 billion U.S. auto insurance industry. 

Driven by severe technician shortages, rising labor costs, repair complexities related to in-car technologies, and parts delays, the average cost of repair has more than doubled since 2019. 

Independent body shops are declining in number as VC-backed multi-shop operators (MSOs) proliferate. MSOs are raising rates and limiting their participation in insurer direct-repair networks. 

Capital formation in legal services and changing public sentiment on justice and fair settlement have increased the number of insurance claims going to litigation by 47% over the last three years. 

Medical costs show no signs of slowing, and new Medicare Secondary Payer rules will shift more costs to auto insurers moving forward.

These developments feel more structural than cyclical. The same can be said for the homeowners and broader commercial markets, where climate perils loom and new species of cyber losses (e.g., CrowdStrike) evolve as fast as technology itself.   

Insurance leaders are asking themselves the same questions they always have:

• Are our people working on the right things?
• Are our people working to the quality standard that’s needed?
• How fast are we cycling, operating, executing?
• How fast are we innovating? Moving with the market?
• What are our real costs related to all this?

What feels different these days is the urgency to find answers--concrete, quantifiable answers, in real time. Everyone has an opinion depending on their department, their patch. The goal is to generate a single source of operational truth based on how people are really working. 

A new crop of software has emerged generating valuable operational intelligence without having to master query languages, statistical modeling, or the command line. Offerings from the likes of Skan, Mimica, Fortress IQ, Soroco, Automat, and Arkifi require minimal involvement from the IT department to install and create zero technology dependencies.  

Installing this software on worker machines is way more philosophical and cultural than it is technical. The point to make is that it seems increasingly essential. As a baseball leadership team shunning player and positioning analytics operates at a disadvantage, so, too, an insurance leadership team.  

There are three modes of implementation with this stuff. The first is load, don’t tell. Most employment agreements permit this. The second mode is load and tell--that is, let workers know it’s there, period. The third mode is load and share. Again, the baseball comparison is apt. Real metrics ranking team members published in real time can be a powerful motivator.    

In the now, meaning the current quarter, the controllable input metrics generated by these tools are incredibly helpful driving efficiency and optimizing for, typically, cost. Here are examples in three key areas, working customer-back:

And then there’s enterprise AI. Most insurers are rightly reluctant to use open frontier models like ChatGPT and Anthropic’s Claude to do real work. Many carriers are moving in the direction of building custom models trained on their own data. Operational intelligence tools build a “digital twin” of the humans and systems they’re observing, useful for model training whenever it comes. Scaling laws state: the more data, the better the AI. 

When the board asks, “What are we doing in AI?”, “building a digital twin of our operation for eventual in-house proprietary AI training” is never a bad answer. Especially when that same data drives operational gains paying for its creation.  

Even as large language models (LLMs) have made their way into almost every new mainstream product, some industry sectors have been slow to adopt AI. Risk management is one of them. Fortunately, that is starting to change.

According to a 2023 Deloitte study, only 1.3% of insurance companies had invested in AI. But data from this year indicates a shift is underway. In Conning’s 2024 survey, 77% of respondents indicated that they are in some stage of adopting AI somewhere within their value chain. This may sound a bit nebulous — some stage, somewhere — but it represents a sizable jump from the 61% of respondents the prior year. Additionally, 67% of insurance companies disclosed they are piloting LLMs.

We also found evidence of change with our own survey, one specifically tailored to risk management. We polled 1,000 risk managers and almost half indicated that they expect to adopt generative AI (GenAI) tools within the next three years. These increases collectively speak to the potential for AI to affect our industry.

Before we look at where we could go with AI, let’s look at the challenges that have slowed adoption in risk management.

See also: Providing Coverage for AI May Be Huge Opportunity

What’s Holding Us Back?

Based on our years of work developing AI systems for insurance and researching the needs within organizations, we’ve identified a few key issues:

• At the organizational level: Change is hard, and many companies struggle with integrating AI technologies into their existing infrastructure. Risk management systems often involve complex legacy systems that are not easily compatible with outside tools. They also may not have a clear line of sight into exactly what they want AI to accomplish or how to measure it. Seamless integration requires a lot of forethought and the right partner to ease the transition and ensure that the system is doing what an organization wants or needs.
• At the IT level: The technical complexity of AI tools requires specialized knowledge and skills. In risk management, as with many sectors, there can be a shortage of in-house expertise to manage and leverage new technologies effectively. Companies need to carefully consider their hiring strategies to account for this or perhaps look into outsourcing options to achieve the best results.
• At the adjuster or claims rep level: One of the big impediments has been that many people still don’t know exactly what AI does, how it works, or what its limitations are. Couple this lack of knowledge with mainstream messages about privacy concerns or AI taking people’s jobs, and hesitancy is understandable. Comprehensive education and training programs would help overcome the hurdle in terms of grasping AI’s potential to make their work easier and more efficient while staying within regulatory guidelines. Workers will also become more comfortable as they increasingly use models like ChatGPT and Gemini in their personal lives. They will see that AI is not intended to replace humans but rather augment their capabilities, arming them with unprecedented insights they can use alongside their own experience and judgment for the best outcomes.

Why Do We Want AI?

With so many considerations, it raises the question: What exactly can AI do for risk management that is so great? A lot! So much, in fact, that the technology can no longer be ignored. Here are some of the biggest and most successful active use cases we see:

• Claims processing: AI is being used to automate tasks like reviewing medical records and legal documents, streamlining the claims process, and reducing errors. Not only is it making claims adjusters more efficient, but it is also removing the drudgery and manual processes of their work so they can concentrate on what matters most.
• Fraud detection: We all know fraud is a big problem in claims. AI algorithms can analyze historical data to identify patterns that spot potentially fraudulent items, alerting adjusters and triggering a human review. Identifying and eliminating fraud, especially early in the claims process, can save organizations millions of dollars each year.
• Reserving: On the flip side, some claims simply require more. AI systems can predict potential cost overruns for these complex cases. This allows risk managers to adjust reserves and prioritize cases accordingly so claims don’t run up any more or last any longer than necessary.
• Underwriting: AI can also be leveraged to analyze vast amounts of patient data — far more than a human could possibly consume — to create more accurate risk profiles almost instantly. With better risk profiles, costs go down. This leads to fairer and more competitive pricing for healthcare policies, which is, in turn, an enormous benefit to patients.
• Knowledge transfer: It’s no secret that risk managers and adjusters are aging out. There will be a massive talent drain when they retire, taking institutional knowledge with them. AI-powered tools can capture their expertise, preserving that invaluable knowledge for future generations.

See also: Cautionary Tales on AI

The Road Ahead

Embracing AI presents an opportunity to advance our industry through improved efficiency, reduced costs, and ultimately, better patient care. These outcomes are worth the effort spent to overcome challenges. The time is right to move the industry forward and to usher in a modern era of risk management.

The insurance sector has been looking for ways to fast-track claims processing to foster better customer experience and maintain revenues. Adopting technology like AI is not just a trend but a must.

See also: Insurance: An Industry Embracing AI

Let’s understand how AI streamlines each step-in claims processing:

1.  Empowering Decision Makers

With AI entering claim processing, the ability to evaluate claims has leveled up. AI can make judgments based on large historical datasets and flag fraudulent activities missed in manual investigations.

2.  Speeding Processing

AI-powered systems drastically speed up processes by automating most of the manual tasks. AI can tackle claims at initial levels and categorize them based on urgency. It can further escalate the conversations to the right authorities in real time. 

Let’s now explore some of the top advantages of claims automation:

1.  Smooth Data Analysis

Replacing manual processes with automation leads to higher employee morale. Automation addresses challenges like miscommunications between departments and other delays in claim resolutions. Automation reduces manual data entry, while processing and analyzing unstructured data.

2.  Customer Satisfaction  

Customer satisfaction and loyalty are closely tied to the speed and accuracy of claims processing. Automated workflows enable insurers to deliver faster resolutions, which translates to happier customers. 

For example, with automation, customers can access the real-time status of their claims via chatbots, while freeing agents to spend time on developing meaningful relationship. McKinsey has found that using the right technology can increase customer satisfaction by 20%. 

3. Compliance

AI can be used to spot potential problems that escape human scrutiny and enhance regulatory compliance.

See also: Cautionary Tales on AI

A Real World Example

A national healthcare insurance provider faced challenges in handling claim errors and corrections. Their system required human intervention for a few items. Subject matter experts (SMEs) reviewed emails, created correct copies, and resubmitted, requiring significant time and resources while adding complexity.

A claims automation solution saved over 1,300 monthly work hours for local claims and reduced processing time to 2.5 minutes per claim. For ITS claims, the system saved over 1,600 monthly work hours, with processing time cut to just one minute.

The robotic process automation (RPA) solution also enhanced accuracy by automating the error correction process.. It deployed bots to read and validate claim data, identify errors, and automatically correct them based on predefined business rules. The solution integrated seamlessly with the company’s existing claims management system.

Future Expansion 

As AI and automation evolve, their impact on the insurance industry is expected to grow.

1.  Advanced AI Models  

AI will likely lead to even more sophisticated models. These models will be capable of understanding nuanced contexts and handling complex claims, such as assessing damage from natural disasters or predicting long-term healthcare costs. 

2.  Integration With IoT and Telematics 

Insurers can gain insights into risk factors with real-time data collection. This allows for more accurate assessments and personalized policies.

3. Ethical AI and Transparency  

There will be a growing emphasis on ethical AI and transparency. Insurers must ensure their AI systems are fair, unbiased, and explainable. This will maintain trust with customers and regulators as AI becomes more integral.

4. Expansion of Automation Across Insurance Functions  

Beyond claims processing, automation will expand into underwriting, customer service, and policy administration. This holistic approach will enable insurers to create more efficient and customer-centric operations. 

AI technology is set to create endless opportunities in insurance.

If you've watched "The Good Place" — and you should, if you haven't already — you saw an enactment of a deep philosophical question known as "the trolley problem." The notion is that you're on a trolley heading down a hill, and the brakes fail. If you keep going straight, you're going to kill five people. You can throw a switch and head off onto a siding, but then you're going to kill one person. 

Do you save the five people and accept the responsibility of killing someone? Does your thinking change if that one person is a friend of yours?

The trolley problem may seem like an odd one to include in a comedy, but the handling is extremely funny, and, of course, no character stays dead. 

And the problem neatly exemplifies one of the issues that companies will face as they roll out AI that touches customers directly. If, like Chidi, the moral philosopher in "The Good Place," you make a spur-of-the-moment decision, you are given some grace because you're only human and can only process so fast. But AI doesn't get that grace. Someone sat down ahead of time and programmed or at least developed the AI, so whatever decision is made is treated as well-thought-out — and has to be defended.

AI is held to a much higher standard than we humans are. You can't just decide your AI is good to go once it outperforms your current approach when dealing with customers. You have to account for what people expect out of AI. 

The higher standards for AI have shown up recently in a spate of articles complaining about drones that use computer vision to inspect roofs. The technology sometimes says there is moss or some other problem that warrants denial of coverage when there is, in fact, no problem. 

The systems already do a better job than could be accomplished by having a host of inspectors climb ladders and tromp around on roofs, but homeowners aren't using the current system as their benchmark. They've been led to believe that computers are nearly infallible and that AI is close to magic, so they don't tolerate errors — and often complain to reporters, who share many of those attitudes and are happy to ding AI when it messes up. 

Phil Koopman, a professor at Carnegie Mellon who has a popular newsletter on driverless cars, writes: "It’s simple: people over-trust too soon, and backlash too hard just as quickly after adverse news."

While nearly 41,000 people died in accidents on U.S. roads last year, I'd bet that none got as much attention as the non-fatal accident involving a Cruise autonomous vehicle. The accident was gruesome: Although the AV was initially blameless, hitting a jaywalking pedestrian who was tossed into its path when another car hit her, the AV then pulled off to the side of the road, unaware that the pedestrian was underneath the car and was being dragged 20 feet. But the involvement of the AI greatly heightened the industry and the willingness to find blame — among other repercussions, the CEO of Cruise lost his job, and Cruise lost its license to operate autonomous robotaxis in San Francisco.

My second caution about AI is that, as an article in The Byte expresses it: 

"So-called 'automated decision-making' is being heralded as the next big thing  — but it turns out that many consumers are disgusted by the idea of AI making choices for them."

The article cites a survey that isn't specifically about insurance; it's about job hiring, banking, renting, medical diagnoses, and surveillance. But it's pretty easy to extend the survey results apply to insurance. 

Just as customers want a human making decisions on their loan or job applications, I'd bet that customers don't want to be told that they were denied coverage by AI or had a claim lowered or denied by AI.

AI will increasingly be used to make decisions that touch clients — as it should be — but, at least for now, I'd suggest having humans make the final call and communicate those decisions.

You'll still be required to defend the AI's role in decisions, and people won't give you the benefit of the doubt that they might give to a human acting under time and other pressures.

But at least you won't have to deal with the theatrics that the Ted Danson character summoned for poor Chidi.

Cheers,

Paul

P.S. Here is the trolley scene from "The Good Place." Watching it will be the best three minutes of your day.

The US P&C insurance industry faces many challenges as it struggles to get a handle on a changing economic, social and climate landscape. The list of perils seem to grow every year from the increasing frequency and severity of catastrophic and other weather related events to social inflation, population shifts and regulatory rate hurdles to expanding competitive pressures.

Carriers have little control over these external forces. Typically, they resort to using the conventional levers including extensive rate increases, dramatic changes to eligibility requirements and product changes that eliminate coverage. While these approaches can alleviate some carrier issues, the new truth is that they are not enough to make a positive impact for the business.

The industry is rife with stories of carriers going insolvent, soaring reinsurance costs and entire sections of the US having unprecedented struggles accessing basic insurance products. US insurers are in need of forward-thinking, data-driven solutions to help them grow profitably all while achieving adequate, fair underwriting practices and increasing accessibility. 

One potential solution to the problem is advancing our knowledge of individual insureds and their needs. 

Understanding customer behavior and risk tendencies enables insurers to better predict outcomes and make targeted decisions that support profitable growth. Today, most carriers rely on predictive modeling to stay competitive. The industry recognizes the power of analytics in categorizing insureds into broad groups based on factors like credit scores, claim-free periods, and residential zip codes. However, advancements in technology have confirmed that the traditional variables are too broad, too bulky and do not provide the stratification we need to fairly and accurately assign risk.   

Person-level behavioral insights involve a big data approach to evaluating an individual's propensity for some target variable based on their behaviors and actions. By using a vast amount of information including but not limited to a specific insured's interests, purchasing behaviors, media consumption, and associated brands, these newer and powerful AI models are able to more accurately predict frequency, severity, loss and other insurance related outcomes. Insurers will be better equipped with these insights to navigate market volatility, more accurately assign risk, avoid adverse selection, and achieve underwriting profitability and growth.

Of course, there are important questions pertaining to bias when assessing this level of information on an individual. With modern statistical techniques, these models must be built with stopgaps and intense scrutiny on their potentially disparate impacts to protected groups.  Luckily, these techniques exist within the behavioral insight industry and have even been proven to improve the biases the insurance industry sees in its existing underwriting processes. 

Overall the goal of integrating behavioral insights into modern insurance is to create a more targeted, profitable, available and less-biased product. As the industry evolves, carriers that embrace behavioral intelligence will be better equipped to face uncertainty, proactively respond to challenges, and secure their position in the market. In an era where the stakes are higher than ever, the ability to accurately assess and manage risk will distinguish the leaders from the laggards in the insurance industry.

About Devyn McNicoll, Pinpoint’s Head of Actuarial

Devyn McNicoll a traditionally trained actuary with over 10 years of experience helping companies grow profitability in the P&C insurance space. She is an Associate of the Casualty Actuarial Society (ACAS), a Certified Specialist in Predictive Analytics (CSPA) and holds her Master’s degree in Statistics from North Carolina State University. She deeply values the role of data analytics and mathematics in the industry. Prior to Pinpoint, she was in an executive leadership role at an MGA offering Homeowners and Commercial Property products. She has extensive experience in actuarial pricing, reserving, filing, modeling and leadership at several large national insurance carriers. In 2023, she was one of two actuaries who represented the US in the Young Actuaries World Cup as a semi-finalist in the competition.

Sponsored by ITL Partner: Pinpoint Predictive

Insurance agents are at the forefront of handling sensitive health-related, personal, and financial data. They are thus prime targets for cybercriminals at a time when cybercrime is expected to cost the global economy more than $10.5 trillion by 2026 and when the global cyber insurance market is expected to soar by nearly a factor of six by 2032, to $120.47 billion. 

Agents must remain attentive; recognizing the scope of their risks strengthens security.

Types of Cybercrimes and Their Prevention

Cyber risks are growing more complex. Here are typical forms of cybercrime and how you can safeguard your agency.

1. Malware

Malware, or malicious software, intends to penetrate and destroy systems while frequently stealing important data. It can take many forms, including viruses, worms, and trojans. Email attachments or USB devices can spread it.

Prevention Tips:

• Install and update antivirus software: Do not forget to renew your antivirus software because it is always the first step to detecting new malware.
• Educate your team: Teach your employees to recognize unwanted emails or connections and avoid downloading unauthentic software files. In addition, ensure spreading awareness about emails with attached files received from an unknown source.
• Regularly back up data: Always back up your data to prevent data loss due to malware infecting your computer.

2. Phishing

Phishing attacks are socially engineered techniques in which an attacker spoofs as a trusted entity. Phishing aims to deceive people into disclosing sensitive information, tempting them with convincing calls to action or offers to click on dangerous websites. These URLs are often distributed via email, SMS, or calendar invitations.

Prevention Tips:

• Check whether the sources of the email are verified: Always check if the sender's address is correct and watch out for odd requests.
• Get multifactor authentication (MFA): Extra security measures prevent unauthorized access, even if the login credentials are exposed.
• Use spam filters: Use enhanced spam filters to block phishing emails before they reach your inbox.

3. Ransomware

Ransomware is software that encrypts files and demands money for their release. Ransomware potentially causes irreversible data loss as well as financial and operational destruction.

Prevention Tips:

• Make regular backups of your data: This creates a safe copy of your data that you can retrieve after ransomware breaches.
• Always keep your software up to date: Secure your operating systems and applications by keeping them up to date. Ensure you fix all the vulnerabilities that ransomware can use.
• Set administrative permissions to the least level: Another best way to keep ransomware from spreading through your network is limiting user privileges.

4. Insider Threats

Insider threats emerge from within an organization. They can be deliberate or unintentional. Any current or former employee, contractor, or business partner can misuse their rights. They can steal data or fall for a phishing scheme, resulting in a cyberattack.

Prevention Tips:

• Set access controls: Enforce policies so employees only access the information necessary for their job. Also, update the roles that people currently hold.
• Administer employee activity: With the monitoring tool, you can discover unusual actions that suggest an imminent insider threat.
• Positive work environment: Having many dissatisfied employees may lead to the likelihood of insiders committing fraudulent activities.

Implement preventive steps and stay current on prevalent cybercrimes to considerably lower insurance agency breaches. Staying up to date on cybersecurity defenses can secure your business and clients from online threats.

Best Cybersecurity Practices for Insurance Agents

Insurance agents must prioritize their clients' information with robust cybersecurity practices:

1. Comprehensive Security Training

Ensure that your employees are frequently trained in how to identify forms of cyber threats and how to handle them. This will provide the first level of protection against attacks and makes most tasks within your agency secure.

2. Using Digital Certificates

Insurance agents should use SSL certificates to keep clients' sensitive data secure. These data include names, addresses, emails, and Social Security numbers. SSL certificates create a secure layer to share data between the user and server.

Using HTTPS maintains your business reputation. When an insurance company or website does not use an SSL certificate, Google flags that as a non-secure site, which can affect your brand identity.

There are various types of security certificates based on the needs of businesses. Companies can choose from multiple certificates like domain validation, wildcard SSL certificate, and extended validation.

3. Multifactor Authentication (MFA)

Implementations relying on or only using passwords for the authorization of the systems are not solid enough. Integrating MFA takes security to the next level and can be done simply:

• Login Verification: Ask your employees and clients to enter a secondary code sent to their mobile devices after they enter their passwords, or have them unlock their devices with a fingerprint. 
• Hardware Tokens: High-level employees who might have to deal with sensitive data can use hardware tokens that generate time-sensitive codes that allow access.

4. Regular Data Backups

Backing up should be considered in the same way as an insurance policy for your data. Backing up relevant data often enough means that you can get back into operation in a short time after a cyberattack or failure of IT systems. When it comes to backups, go for several tiers of backup, such as local, offsite and cloud-based.

Types of Backups:

• Local Backups: Use external hard drives or NAS devices.
• Offsite Backups: Consider tape or remote server backups.
• Cloud Backups: Use AWS, Google Cloud, or Microsoft Azure for scalable, automated solutions.

5. Controlled Access to Data

Not every employee working in your agency requires access to all the data available with the organization. It is necessary to enforce information security policies that would limit the accessibility of specific data to a few individuals. This helps reduce the harm coming from internal threats and minimizes the possibilities of a leak of the firm's information.

6. Secure Communication Channels

Security will always recommend the use of encryption when it comes to communications. Make sure that all data are encrypted, while they are being transferred, as well as during storage. This helps to eliminate instances where third parties may be intercepting and decoding your information.

7. Engage Security Professionals

It could also be wise to seek help from professionals as a viable strategy. Whether you hire your security team or consult cybersecurity specialists, it is always beneficial to have someone knowledgeable guide you when assessing and addressing risks.

8. Network Traffic Monitoring

Surveillance of the network must be done keenly. Network monitoring lets you identify suspicious activities likely associated with a cyberattack. It is also important for organizations to learn to prevent a breach from happening in the first place or at least provide early alerts.

9. Incident Response Planning

Even with the best efforts, you can still find yourself facing a breach. In the event of a security incident, your team must have a clear copy of the response plan. This can greatly reduce the time it takes to recover from an attack and minimize data loss.

These security measures effectively safeguard sensitive data and protect your agency.