It's Not Human Vs. Machine; It's PLUS

What is right for your business, and where do you strike the balance between people-focused and automated processes?


When it comes to thinking about how to secure data assets, deploy analytics and deliver intelligent automation, the question of "why are we doing this?" is well worth thinking through carefully. What is right for your business, and where do you strike the balance between people-focused and automated processes?

Those questions are the challenge underlying intelligent intervention. While the specifics of the answers will obviously vary by company, we believe they involve a common objective – to deploy the right resources to the situation at hand, whether that be full automation without underwriter or claims handler involvement, or a subject matter expert making the call based on insightful decision support.

The importance of data strategy

Any aspirations for achieving that objective will rely on granular data, and a culture that recognizes the power of data to both fully automate where it makes sense and provide critical decision support for underwriters and claims experts where it doesn’t.

Consequently, a data strategy will need to deliver a single, complete, consistent enterprise-wide source of facts relating to risk and non-risk activities. The strategy will also have to serve functional requirements for reporting, interactive dashboarding and data visualization to enable frictionless consumption of facts by key stakeholders/systems.

A suitable environment in terms of infrastructure, architecture, culture and tools will allow testing of new data sources and enable analytics and business teams to test hypotheses. Companies will also have to ensure that roles and responsibilities are clear and that there are internal control, governance and communication structures in place to leverage this data resource in pursuit of a single version of the truth.

Figure 1: A data strategy framework for breaking the cycle of poor data

As for the data itself, our view has long been that big data starts at home, so a lot of the gains can come from a company’s underlying data assets. There is undoubtedly a very large and growing source of invaluable third-party data available, but, without a core data asset to connect to, the value is diminished.

Insurers also have huge volumes of unstructured data, which is an area that will continue to offer significant potential competitive advantage. Improving loss cost analytics with claims reports, creating reusable insights from surveys and adapting claims strategies following court or medical reports are all possible. 

See also: COVID-19 and Need for Analytical Insurers

Automation and AI – a broad church

Moving on to how to apply that data, when companies think today of more automated approaches to underwriting, pricing and claims, either artificial intelligence (AI) or machine learning are often seen as the default solutions.  

To some extent, this stems from the overuse of the term AI or the broad use of the term. In fact, AI can encompass a huge breadth of technology, from very deep predictive models through to cognitive learning. Most of the industry spending to date has been at the predictive modeling end, and, while investment will expand through to the cognitive area as the technology matures, there are still very large gains to be made in claims and underwriting analytics using existing methods and approaches.

Deployment challenges

For companies needing motivation to act in this space, the huge range of challenges and uncertainty in the market - geopolitical and technological uncertainty, given the pace of change, as well as legal and regulation uncertainty – should be more than sufficient. The aim of pricing, underwriting and claims technology investment should be to create the capabilities to navigate this uncertainty.

Deployment challenges vary across insurers, but often manifest as lack of pace and agility. Insurers need an environment that enables technology and data to be used at pace to make great decisions at both product and portfolio level and link their technical and business communities. This will help ensure a rich seam of insight running backward and forward between internal communities.

Some companies were able to use technology and data to respond quickly to changes in the market. A recent example is the FSA announcements on new and renewal pricing. Leading insurers have been able to resolve their strategy and deploy adjusted rates inside a matter of days. Others have taken months. Equally, companies in the first group could be far more surgical in their approach and achieve better outcomes. 

COVID-19 only reinforces these challenges, but on a very large scale, presenting competitive opportunity for companies that are able to quickly adapt their underwriting rules or perhaps change their automated footprint.

Business-wide approach

An important point to recognize in intelligent intervention is that, important as it is, the technology itself is not the be all and end all. Intelligent use of automation and analytics is not a siloed activity and still, most certainly, involves people.

Some high-level challenges will need the time and attention of the business as a whole.

How will new technologies integrate with existing systems? Legacy systems are a fact of life for most (re)insurers and will often represent years of investment. Careful thought needs to be given to how to promote connectivity when implementing automation and decision support, including the benefits of capturing both structured and unstructured data, so that people within the business and customers who need information have it, when they want it.

Also, do we have the people and skills to make automation work for us? Not only will you probably need to tap in to a new and different talent stream, but you may be asking your existing people, such as underwriters and claims handlers, to work in new ways. Business culture, training, skills and career development, working practices and reward structures will all possibly need reviewing. 

Connected specialisms

The point is that automation in (re)insurance, or the use of automation for intelligent intervention at least, needs a connected, business-wide approach, albeit with a technological flavor. 

For example, some insurers already recognize that being highly effective at claims estimating opens up opportunities for their portfolio management, underwriting and actuarial teams. Reliable claims estimating provides these functions with a sound basis on which they can confidently make day-to-day business decisions, enabling them to be first movers in a market or agile in changing circumstances, both vital attributes in a highly competitive environment. The digitization of data assets delivers improved sophistication but also reduces frictional costs.  

See also: 7 ‘Laws of Zero’ Will Shape Future

Asking the right questions

Albert Einstein said, “The measure of intelligence is the ability to change.” It’s no coincidence that, in our experience, companies that do analytics and automation well and with a clear strategy linking pricing, underwriting and claims are best placed to use them to initiate intelligent interventions. The flexibility, agility, speed to market and cost savings that typically come with them naturally deliver benefits from both effectiveness and operational efficiency. 

That’s why, when it comes to analytics and automation technology, asking the right questions and addressing the right challenges and issues is critical. Pursuing loss ratio benefits, essentially a measure of effectiveness, will deliver efficiency savings, too.

Dave Ovenden

Profile picture for user DaveOvenden

Dave Ovenden

Dave Ovenden is the global lead of pricing, product, claims and underwriting at Willis Towers Watson’s insurance consulting and technology business.


Read More