Insurance provides compensation to policyholders against the losses resulting from specifically identified risks. ‘Risk’ in insurance terms refers to the likelihood of adverse events that will hinder the policy holder’s specified objectives.
There has been a continuous and steady increase in business insurance premiums. Global commercial insurance prices increased by 15% in 2020, marking an increase for the 15th consecutive quarter. Additionally, global premium volumes are forecasted to surpass USD 7 trillion for the first time by the end of 2022, facilitated by economic recovery after the pandemic. During COVID-19, social restrictions reduced risk exposure from businesses across industries, including hospitality and F&B, which has decelerated the rate of premium increases. However, the growth is likely to resume as economies gradually recover.
This post will consider the factors that affect insurance premiums and how insurers and businesses can use AI to control the cost of insurance premiums through enhanced risk management processes to prevent and mitigate legal claims
Insurance coverage common across many industries includes insurance for public liability, employers’ liability, professional indemnity, and buildings and contents insurance. Simultaneously, some coverage is industry-specific, such as insurance for construction and engineering, commercial property, commercial motor and marine.
The insurance premium charged to the policyholder is a factor of (1) The client company’s profile; (2) The insurer’s internal expenses; (3) The insurer’s profit margin required to generate investment returns; (4) Intermediary charges (if any).
The client profile considers the (1) Nature of the business to be insured; (2) Annual turnover of the business; (3) Number of people employed; and (4) the business’ insurance claims history. When considering the claims history, businesses usually fall into the category of being (a) Short-tailed in which claims are settled quickly; or (b) Long-tailed, where claims are usually more complex and have a more extended settlement period until a length of time beyond the policy year.
Factors that affect insurance premiums
The primary methods used for the calculation of insurance premiums include (1) Experience rating or (2) Exposure rating, or a combination of both depending on portfolio level factors. For example, calculations for SMEs or younger companies may be weighted towards exposure rating due to limited historical data. In contrast, estimates for larger companies with more data points may weigh towards experience rating.
For example, the experience rating for professional indemnity insurance will consider the history of claims received against the company, its frequency and severity. On the other hand, exposure ratings are often industry-specific. For example, within the construction industry, key exposure may relate to the size and safety of a proposed project. Exposure could also result from changing regulatory attitudes, as seen with product liability claims discussed in our blog post, where attitudes have become more claimant-friendly.
Improving risk management with AI
The use of AI risk management tools by both insurers and policyholders can assist with controlling the cost of insurance premiums by enhancing (1) The accuracy of risk assessment; and (2) Risk management to reduce the realisation of risk.
For the insurer, risk pricing determines the premium, based on the likelihood of the insured making a claim and the probable size of the claim. There are two potential improvements that AI can contribute to the pricing stage.
This includes (1) Data veracity, by increasing the accuracy of data models to enhance decision making; (2) Enhanced understanding of client preferences and habits by incorporating historical data relating to the company’s profile with greater efficiency to assess the likelihood of risk, and also allowing insurers to set insurance prices more competitively.
Importantly, AI can play a filtering role to assist in fraud detection. The cost of insurance fraud is estimated at over USD 40 billion per year , which incentivises insurers to increase premium estimates by factoring in the risk of fraud. Premium diversion is an example for businesses that purchase insurance via an intermediary. Instead of sending premium payments to the insurer, payments are kept for the intermediary’s personal use through making claims. The use of AI can assist in detecting abnormalities in claim data to reduce the likelihood of fraudulent behaviour.
For the policyholder, there is also an incentive to reduce risk exposure. Insurance premiums are generally lowered where the policyholder is less likely to make a claim. Cost savings can be achieved by enhancing the business’s ability to accurately anticipate and measure risk and take appropriate actions to reduce losses. This can be seen from two use-cases regarding professional indemnity insurance and employer’s liability insurance.
Professional indemnity insurance covers legal defence costs and compensatory damages when the policyholder faces negligent professional advice or service claims. Coverage also includes civil liability claims arising from breach of duty or unintentional infringement of intellectual property. This coverage is essential for professional service providers, industries where disputes over copyright and product quality are common, and businesses with access to confidential customer information. A company with a history of professional indemnity claims may be perceived by potential insurers as not having adequate risk management protocols. In particular, if many of these claims involve long-tail liability, a higher insurance premium might result in the future.
Natural language processing (NLP) can be used to reduce the occurrence of claims by addressing potential risks early. For example, the ability of NLP to classify and analyse natural human language can help to identify nuances within correspondences between a business and its clients to detect the risk of legal claims. As we discuss in an earlier blogpost on ‘Customer Dispute Prevention in the Financial Sector’, if a customer complaint escalates into a potential dispute, effective deployment of NLP in a business' risk management system can help to ascertain the urgency of such a complaint and promptly direct the complaint to suitable personnel.
NLP can be applied to different forms of communications, including feedback forms, social media activity and call transcripts to identify the root causes of customer dissatisfaction. Customer feedback is particularly important for professional indemnity insurance, where disputes relating to the quality of service can escalate into claims for negligence or misrepresentation.
Employers' liability insurance covers the policyholder against compensation for employees’ work-related injuries or illness claims. In the UK, for example, such coverage is compulsory for employers with full and part-time employees. Premiums can range from £60 to over £210 per worker annually, depending on the size and nature of the business. As we observed in our article on ‘Reducing workplace conflicts with AI-powered tools’, around 874, 000 employees are estimated to take sick leave as a result of workplace conflict each year, at an estimated cost of £2.2 billion to employers in the UK. The prevalence of this type of work-related illnesses can give rise to a significant number of claims against the employer, which can translate into increased insurance premiums.
In the employment context, NLP tools can be used to identify risks of workplace harassment. A study modelled to predict harassment in a social media setting achieved an accuracy rate of 92.8% . The process involves (1) Pre-processing to filter keywords; (2) Features extraction where text is transformed into algorithms amenable to the programme; (3) Classification where extracted features are integrated into the classifier to reach predictions . These predictors can similarly be integrated into other communications systems such as emails, Teams or Slack to detect risks of workplace harassment and minimise the occurrence of related legal claims.
Enhancing the ability for insurers and policyholders to more accurately assess risk and mitigate the occurrence risk with risk management will be essential for businesses to control insurance premiums. AI creates such an opportunity, which can help insurers maintain competitive pricing in the market as well as for policyholders to redirect financial resources to initiatives that will drive business growth and profitability.