On demand and usage based insurance

The combination of ‘Big Data’ in conjunction with technological advances in artificial intelligence (AI), predictive analytics and blockchain create the foundation and operational capacity for new insurance products and opens doors to new and exciting opportunities within the insurance industry.

This certainly is the case with “on-demand” insurance, a term used often interchangeably with “usage-based” insurance. There are differences between the various types of “on-demand” insurance and the term ‘on-demand’ is itself open to various interpretations: Scott Walchek, founding chairman and CEO of pioneering on-demand insurance platform Trōv, described it as “giving people agency over the items they own and enabling them to turn on insurance cover whenever they want for whatever they want — often for just a single item.”¹

In its “purest” sense, on-demand insurance describes a situation where an insured activates their coverage by way of a smart device or application, or the cover is automatically based on criteria such as location, activity or context.  The cover, similarly, is terminated manually or automatically and the customer can choose to turn on insurance from different providers at different moments.

In usage-based insurance, the connectivity between the customer or insured and the insurer is more substantial as the contractual relationship between insured and insurance company persists continuously (that is, also in times where the item is not “in use”).  The usage-based component is often part of a traditional insurance coverage. A prominent example is usage-based insurance for cars, where basic liability and theft insurance may be always in force and additional premiums for liability and collision cover are calculated upon usage (for example, based on distance travelled, countries visited or driving habits).

Research published by the International Underwriting Association (IUA) observes that “pay-as-you-go” models of cover will allow customers automatically to activate policies when and where they need them.² Examples include on-demand insurance being available for valuable personal possessions, drones, motor vehicles, homeowners and home sharing hosts, travel and event insurance, small business insurance, insurance offerings for workers in the gig economy, and cover for digital businesses against loss due to employer’s liability, public liability, professional indemnity, cyber liability and directors’ and officers’ liability.

One of the biggest benefits of Big Data combined with AI analysis ,  is that it provides new sources of information and insight for insurers into specific risks and the risks of individuals, groups and types of insureds.  It also provides predictive analytics at an individual insured level, to enable insurers to predict and price risk more accurately (and quickly).  This enables more granular segmentation of risks, increases the effectiveness of risk identification and also allows for pricing that is both quicker and more risk-sensitive.

An example of a usage-based insurance product is Flock Cover’s Connected Motor Fleet Insurance.

Flock’s engine combines telematics and environment data to quantify risk on a per-second and per-metre basis. This data-driven approach to underwriting takes into account driver behaviour, location and decades of crime and accident data, with the aim of providing prices that reflect risk on a per-fleet basis.

Antton Peña, the Founder and Chief Strategy Officer at Flock Cover observes that approach to insurance “allows fleets to be underwritten based on actual exposure to risk rather than just historical data. Flock’s underwriting model continuously learns as more data is ingested allowing it to continue improving as the world becomes increasingly connected and autonomous”.

The use of real-time data is not only of benefit to the insurer but the lessons from such data can be provided to the insured.  In the case of commercial motor fleets, this can allow them to identify, understand, quantify and mitigate risks, incentivising them to do so with lower insurance premiums and rebates (see e.g. Safety Insights). As vehicles and the environment around them become increasingly connected, it is likely that even more data will be capable of being collected in the future to improve the management of risk.

Challenges and risks

On-demand insurance is growing rapidly with predictions that, by 2030, the global insurance market will evolve to contain highly dynamic, usage-based products that are tailored to individual customer behaviors. Alongside this product evolution, it has been predicted that many consumers will transition away from the traditional annual renewal model to continuous on-demand products.³

This convergence of new technologies and consumer demands is creating new and exciting opportunities within the insurance industry, but also complex challenges including the determination of liability for harm or damage, privacy considerations, cyber security risks and insurer solvency.

Actual and prospective impacts upon insurance law and practice, serve to highlight opportunities and risks in navigating the changing or changed landscape.⁴

Regulatory Considerations

In an increasingly globalised and interconnected commercial world, many of the challenges and opportunities presented by on-demand insurance will engage the regulatory frameworks of multiple jurisdictions meaning that the regulatory position must be ascertained in each relevant jurisdiction.  Often, a product will need to be adjusted in different ways for different jurisdictions, to meet the particular requirements in each of those jurisdictions.

Emerging issues related to insurers or licensees leveraging new technologies to develop products for on-demand insurance purposes can have global implications and impacts – including, but not limited to, reviewing new products, cancellations, non-renewals, coverage issues, notice provisions and policy delivery requirements.

The novel features of products such as on-demand insurance were not always envisaged at the time of existing regulations being implemented.  As a result, regulations do not always cater for the features of such technological advances, making it difficult to determine how certain regulations apply to those features.  For example, requirements to provide information prior to inception of a policy could be interpreted (depending on the rules in question and the exact features of the relevant product) as requiring that information to be provided each time cover is switched on. Therefore, parties designing and launching new products in this space need to take particular care to ensure that their products comply with regulatory requirements.

Privacy

Greater access to insureds’ data and activities raises risks from inadvertent data misuse right up to data breach and cyber fraud.  This might involve insurers obtaining unauthorised knowledge of facets of consumers ́ lives, including sensitive data concerning the customer’s health, location, or financial status. Accordingly, insurers will have to consider very carefully how their activities comply with data protection law (or rather multiple different data regimes across multiple jurisdictions).

Disclosure and fraud

On-demand insurance poses new challenges for insurers, as the window for detecting and stopping fraudsters is truncated. This is not particularly a problem with an on-demand usage-based insurance model, where a contractual relationship between insured and insurance company persists continuously (i.e. also in times where the item is not “in use”).

Conversely with short-term, ad hoc insurance coverage, the National Association of Insurance Commissioners observe “since coverage can be turned on and off easily with a swipe on a smartphone, the possibility of fraud risks increases with consumers who only turn on their insurance when wanting to make a claim.” ⁵

However, as delivery of on-demand insurance is typically data-driven, and some products are even sensor-based (such as telematics in a car), the opportunity to commit fraud is in principle lessened – for example, insureds do not have the opportunity to give (whether deliberately or otherwise) inaccurate answers to questions in a proposal form. Information asymmetry will also arguably decrease as the use of technology powering on-demand offers matures. This would allow insurers to charge higher premiums to individuals who present higher risks. Such differentiation would, in theory, drive out high-risk individuals as premiums for them would skyrocket.

Moreover, as Ernst & Young⁶ argue, distributed ledger technology or blockchain can have a major impact upon fraud detection and risk prevention due to its ability to provide a public ledger across multiple untrusted parties, thereby delivering the potential to eliminate errors and detect fraudulent activity.

Of course, care also has to be taken with AI to ensure that models do not mistakenly determine that certain behaviours, in particular by vulnerable customers, is fraudulent when in reality there is a reasonable explanation for that behaviour. To date, Courts and regulators have generally sought to make companies which use AI responsible for that use, rather than allowing the companies to hold the AI at arm’s length and decline liability for its acts.

Concluding comments

There is no doubt that the new on-demand insurance products that are emerging will play a fundamental role in the future of the insurance industry generally.

On demand and usage-based products not only change the way insurance is established and delivered, but their technological foundations and connectivity create significant avenues and opportunities for risk mitigation and management.

Equally, these products present challenges, in particular regarding their regulatory treatment, so great care must be taken when designing and launching them.

The article was co-authored by Dr Anthony Tarr, a Senior Consultant.

Footnotes

1. See Graham Buck, “Kiss Your Annual Renewal Goodbye; On-Demand Insurance Challenges the Traditional Policy” 14 September 2018. https://riskandinsurance.com/on-demand-insurance-challenges-traditional-policy-constraints/
2. www.iua.co.uk/IUA_Member/Press/Press_Releases_2019/IUA_publishes_on-demand_insurance_report.aspx?WebsiteKey=84dca912-b4fb-4a0f-a6e5-47ad899350aa
3. Tanguy Caitlin, “Insurtech-the Threat That Inspires,” McKinsey & Company, 2017, https://www.mckinsey.com/industries/financial-services/our-insights/insurtech-the-threat-that-inspires (hereafter McKinsey 2017)
4. Generally, see Julie-Anne Tarr and Anthony Tarr, “On-Demand Insurance and the evolving technological and legal environment”, (2021) Journal of Business Law 535.
5. Center for Insurance Policy and Research, “On-Demand Insurance,” content.naic.org, 5 November 2022, https://content.naic.org/cipr-topics/demand-insurance (hereafter CIPR On-Demand)
6. Ernst & Young, “Blockchain in Insurance: applications and pursuing a path to adoption,” https://www.ey.com/Publication/vwLUAssets/EY-blockhain-in-insurance/$FILE/EY-blockhain-in-insurance.pdf ‘A decentralized digital repository can independently verify the authenticity of customers, policies and transactions (such as claims) by providing a complete historical record. As such insurers would be able to identify duplicate transactions and those involving suspicious parties.’