The last two years have been a reality check for the many insurers that fell short of their aspirations for implementing advanced analytics to improve key areas like customer experience, claims and telematics, says a new report.

The U.S. and Canadian property/casualty insurers surveyed revealed they were either overly ambitious in their goals or were hindered by the realities of day-to-day business and market challenges, according to Willis Towers Watson’s P&C Insurance Advanced Analytics Survey, which compared results to a prior report on the topic from 2017.

Respondents expected they would be further along in using advanced analytics for enhancing customer centricity. For example, 77 percent expected to be using internal customer data as a source by now, but only 54 percent are. Similar gaps in expected versus actual usage have developed for social media (46 percent versus 26 percent) and clickstream data (34 percent versus 14 percent). While most carriers are still optimistic about the future, just over 10 percent of respondents said they are not trying to improve customer centricity.

When it comes to claims management, only commercial lines insurers are ahead of schedule: 53 percent are utilizing unstructured internal claim information as a data source compared to 41 percent two years ago. The usage in claims varies by line of business: 54 percent of workers comp writers said they used advanced analytics in claims compared to 10 percent in commercial property. In personal lines, the use of images (most typically used to verify and expedite claims) has more than doubled to 24 percent of respondents—nearly on par with expectations from two years ago.

Insurer use of telematics data showed relatively slow growth for both personal auto (28 percent to 30 percent) and homeowners (2 percent to 9 percent) over the last two years, and many carriers said they have no plans to use telematics in the next five years.

What’s keeping insurers from meeting their advanced analytics goals? Nearly half (45 percent) of respondents said their progress has been hindered by information bottlenecks where people and systems typically need to interact. They also cited handling infrastructure constraints (37 percent) and a lack of sufficient staff to analyze data (37 percent) as top obstacles preventing them from becoming more data-driven.

Other survey highlights:

• Insurer use of artificial intelligence and machine learning continue to trend upward. The percentage of respondents saying they use both to build risk models for decision-making (26 percent) and to reduce manual input (22 percent) doubled in the last two years, and 60 percent of companies are targeting such capabilities by 2021.

• While many InsurTechs are now looking to collaborate rather than compete with incumbent insurers, integration is still in the early stages for the majority (69 percent) of respondents, and a significant number (22 percent) said they’re not doing anything in this area.

Willis Towers Watson’s 2019 Advanced Analytics Survey asked P/C insurance executives in the U.S. and Canada for their insights on the future of advanced analytics relative to where we are today. Ninety-nine P/C insurers participated in two web-based surveys conducted from April to July 2019. There were 66 multiline carriers, 23 commercial and 10 personal line carriers. Respondents included eight of the top 20 U.S. P&C insurers and eight of the top 20 Canadian P&C insurers.