With more and more being written about AI — and with more and more applications being developed for its use — we couldn’t help wondering about the role of AI in what we’ve come to consider business intelligence (BI). Obviously enough, BI is synonymous with data analytics and has been for a while. But we really wanted to know about BI’s origins and its evolution, which would help us to understand and to keep pace with its evolution with the introduction of AI.

One of the first things we came across was a post from DATAVERSITY, a producer of educational resources for business and Information Technology (IT) professionals on the uses and management of data. By way of historical perspective, DATAVERSITY offered this:

In 1865, Richard Millar Devens presented the phrase “Business Intelligence” (BI) in the “Cyclopædia of Commercial and Business Anecdotes.” He used it to describe how Sir Henry Furnese, a banker, profited from information by gathering and acting on it before his competition … in 1958, an article was written by an IBM computer scientist named Hans Peter Luhn, describing the potential of gathering business intelligence (BI) through the use of technology … In 1968, only individuals with extremely specialized skills could translate data into usable information. At this time, data from multiple sources was normally stored in silos, and research was typically presented in a fragmented, disjointed report that was open to interpretation. Edgar Codd recognized this as a problem, and published a paper in 1970, altering how people thought about databases. His proposal of developing a “relational database model” gained tremendous popularity and was adopted worldwid … The number of BI vendors grew in the 1980s, as business people discovered the value of business intelligence. An assortment of tools was developed during this time, to access and organize data in simpler ways. OLAP [online analytical processing], executive information systems, and data warehouses were some of the tools developed.

Given what we know about the analytical and predictive abilities of AI, it’s fairly easy to generalize, then, about the ways in which it will contribute to the evolution of AI.

Here’s What We Think

While we can’t be sure how much of this has actually come to fruition, it seems safe to assume AI helps or will help BI to:

1. Automatically analyze large datasets, identify patterns, and generate insights.
2. Forecast future trends (given #1), optimize operations, and allow even better data-driven decisions.
3. Interact with BI systems using natural language processing (NLP), making it more accessible to and usable for non-technical people.
4. Use machine-learning algorithms to understand user behavior, adapt to changing business needs, and continuously improve performance.
5. Enable real-time decision-making, making BI dynamically operational, as opposed to reporting-centric.
6. Integrate with other technologies, such as big data, the cloud, and IoT to leverage a wide range of data sources and create unified views of businesses.

Are we correct about all of that? We don’t know. We’d have to combine AI with a crystal ball and a Ouija board to be sure. But we do know we’ve come a long way since 1865. And the evolution of BI will continue with the further adaptation and inclusion of AI.

That’s why we built the Finys Suite to be ready for it.

Your legacy system is starting to wheeze a little. It’s not as reliable as it once was. It’s nowhere near as flexible as it used to seem to be. It requires a little more coaxing than you’d like to give it. And you, your agents, and your policyholders are starting to wonder if it might be ready for a safe spot in the Old Systems Home.

But where do you start?

Well, you start with the facts, of course, beginning here:

1. What deployment models does the vendor offer? Do they provide a choice between SaaS and on-premise platforms? You may have more control on-premise, but it likely will require more internal IT resources. SaaS, on the other hand, offers greater security, scalability, and flexibility. And it may offer lower total cost of ownership (TCO).
2. What’s the vendor’s track record on implementations, migrations, and data conversions? Does it stumble out of the blocks? Does it fade in the late going? Or are its pace and delivery steady from start to finish? Your business operations will depend on the vendor’s performance.
3. How is the process of implementing the software, migrating your data to the new system, and performing the necessary conversions priced? Will you be on the hook for delays or scope creep?
4. What’s the vendor’s approach to change management (process re-engineering and getting your employees up to speed on the new software’s ability to facilitate necessary processes) and training (ensuring buy-in from and a smooth transition for your employees).
5. How good is the software? How much of it is standardized? How much of it is customizable? How reliant do you have to be on the vendor to configure the system, to configure existing products, and to develop new products?

That Was Fun

All five steps above are necessary, and we highly recommend the gathering of as many fact as you need to ensure your decisions about a vendor and its software are fully informed. You can find all manner of data to support the contention that the single most important factor in a property/casualty insurer’s selection of a core processing system is integration with existing systems and data. But it’s not.

The single most important factor in any insurance company’s selection of any software is word of mouth. In other words, the most important thing to establish with a vendor is trust, beginning with the trust it’s established with other insurers. Yes, the software has to be good. But a trusting relationship comes first.

If your legacy system is telling you the search is on, remember this: There is no perfect system. But there is a perfect system for you.

That includes your relationship with the vendor.

The July edition of Best’s Review ran an article called, “Social Inflation Remains a Thorn in the Side of Casualty Insurers”. The article reflects the evolving psychology of some policyholders and the corresponding expectations that yield suspicion of corporations and assumptions about corporations’ abilities to inflate compensatory damages:

Social inflation continues to test the ability of casualty insurers with unpredictable and excessive claim costs … a reflection of shifting social and cultural attitudes toward corporations … when people have claims or file claims … they’re looking at the deep pockets of the corporations and figuring that, “Hey, somebody has to pay for my misfortune” … A lot of that led to an increase in lawsuits … [and] the erosion of tort reform in a number of states.

In other words, we may be facing the proverbial perfect storm of social inflation and nuclear verdicts.

What is Social Inflation?

Social inflation denotes the increase in claim severity above the historical norms of economic inflation and claim trends, in which the rising costs of insurance claims are driven by societal trends and views toward litigation, rather than just general economic inflation. Those trends include changes in public perception and attitudes toward corporations, liability, and risk-taking that can lead to increased litigation and larger jury awards. They include the involvement of outside parties in funding lawsuits that drive up litigation costs. They include reversals of tort-reform measures that were intended to protect insurers from insolvency. And they include varying demographic makeups of jury pools that can influence jury verdicts and awards. The upshot is that those trends lead to increased claim costs, higher premiums, and reduced profitability for insurers.

What is a Nuclear Verdict?

Nuclear verdict denotes verdicts in favor of plaintiffs with damage awards that surpass $10 million. Such verdicts are considered nuclear because they can have devastating effects on defendants, potentially causing financial hardship and bankruptcy. Nuclear verdicts often involve complex cases, such as product liability, medical malpractice, or catastrophic injuries. The increase in nuclear verdicts is attributable to a number of things, including the changing attitudes toward corporations mentioned above, increasingly aggressive plaintiff attorneys, the increasing numbers of class-action lawsuits, the increasing cost of healthcare and medical treatments, and more. The proliferation of nuclear verdicts is a source of concern and consequence for the insurance industry and defense litigators. They lead to increased insurance premiums, reduced coverage options, and a greater risk of financial ruin for defendants. As a result, there is a growing need for effective risk management strategies, litigation tactics, and claim management techniques to mitigate the implications of these verdicts.

Start at the Beginning

We can’t say your core processing suite can save you from all the effects of social inflation and nuclear verdicts. There are regulatory and legal issues to be resolved, as well as social attitudes to be examined and mitigated. But the right suite — one with the flexible configuration capabilities to enable you to anticipate and adapt — will have you better positioned before social inflation goes nuclear.

If you happen to be looking for such a suite, we know some guys.

As artificial intelligence (AI) continues to establish a foothold in insurance, it’s important to remember two things: (1) AI can be invaluable in automating routine tasks to reduce the need for manual intervention. (2) The predictive value of AI lies in hindsight; that is, its ability to analyze data, to identify patterns, and to predict outcomes is retrospective.

Predictive AI combines machine learning, AI, and statistical models to identify relationships between variables and to make predictions. To cite a few examples, AI can:

• Assess and manage risk more effectively by analyzing data from various sources, including social media, credit reports, and medical records.
• Predict the likelihood of claims, enabling insurers to offer more accurate risk assessments and personalized premiums.
• Automate claims processing, reducing the time and effort required to settle claims and improving accuracy.
• Predict customer churn: By analyzing customer behavior and demographics, AI can identify patterns that indicate whether a customer is likely to switch to a competitor, allowing insurers to target them retention efforts.
• Forecast sales: AI can analyze historical sales data, seasonality, and market trends to allow insurers to prioritize the sales or products or lines accordingly.
• Detect anomalies: AI can identify unusual patterns in data, such as unusual policyholder behavior or rating abnormalities, to allow for swift detection and resolution.
• Optimize processes: AI can analyze process data to identify bottlenecks and inefficiencies, enabling insurers to optimize workflows, reduce costs, and increase productivity.

Generative AI, on the other hand, creates new content, such as images, text, and other media, by learning from, aggregating, and adapting existing data patterns. While generative AI is valuable in creative fields and novel problem-solving, its predictive value is limited to generating new content rather than making predictions about future outcomes.

What’s In It For Insurance?

AI is flexing its muscles in the insurance industry, by improving capabilities like customer service, claims processing, underwriting, and fraud detection. Its ability to analyze large datasets and to process information in specifically programmed ways enables insurance companies to enhance customer service by providing personalized policies and improving communication. AI-powered chatbots and virtual assistants can provide 24/7 customer support, answer common questions, and help customers with policy-related inquiries. AI can streamline or eliminate manual, time-consuming tasks; improve the accuracy of rating and underwriting; and bind policies faster. It can aid in policy comparisons, ensuring policyholders get the most suitable products for their needs. And it can help detect and prevent fraud by analyzing patterns and identifying suspicious behavior.

As AI continues to evolve, we can expect it to contribute more significantly to improvements in the insurance industry. Since it relies on data (past) to enable its predictive capabilities (future), it will always look back to look ahead.

In an industry that measures its success retrospectively (based on claims experience), that’s not a bad way to go.

Given the fact that we’re in the business of the mechanics of insurance (processing software), it had never occurred to us to wonder about its psychology … until it did. When we started to indulge our curiosity, one of the first things we found was an article from the December 2019/January 2020 edition of The Actuary called, aptly enough, “Insurance Psychology 101”. The article says this, in part:

It is unlikely the psychologists Dr. Edward Deci and Dr. Richard Ryan were thinking of digital technologies when they introduced their Self-Determination Theory in 1985, prescribing the human need for autonomy, competence and connectedness to sustain a sense of well-being and to flourish … Media psychology is essential to appeasing and understanding the insurance needs of future consumers whose native language is digital and a medium for human connectedness technology.

Well, there’s a mouthful, as well as a mind-full. But what, we wondered, is media psychology?

A Closer Look

Like so many things in actuarial science and psychology, some things are less complex than they appear. According to Dr. Pamela Rutledge, media psychology:

studies the interaction among individuals, groups, society, and technology to help consumers, developers, communicators, and society at large make good decisions … the advent of the Internet and social media has brought it to the fore … The advent of social media has made the landscape more interrelated and complex.

Granting we’re neither actuaries nor psychologists, what we derive from that definition is the apparent fact that consumers, developers, communicators, and society at large expect the Internet and social media to make things easier. What a coincidence. So, do we. That’s why we built the Finys Suite to be flexibly configurable and capable of adapting to emerging insurtech developments.

As for the psychology of insurance, a cynic might say people buy insurance because they’re paranoid. We tend to take more magnanimous view and believe people buy insurance to protect themselves from financial risk. That, too, seems sensibly simple.

Flip a Coin

In the end, it doesn’t really matter why people buy insurance or what media psychology is. What matters is that people do buy insurance, and we support the companies that sell it to them.