Key findings of the study include:
- Many insurers are using conventional machine intelligence (MI) for risk assessment and prediction models, but early adopters of more advanced forms of MI, including machine learning (ML) and artificial intelligence (AI), are benefitting from faster claims settlement, cross selling and risk scoring.
- 80% of insurance companies lack data that's clean and curated enough to implement the most productive machine intelligence. Investing in data engineering is therefore a central factor in yielding greater profit from MI-enabled systems and facilitating deployment at scale.
- This sigma study also found that only 10% of firms are able to implement transformative end-to-end machine intelligence. Insurers therefore need to consider how to holistically employ MI within their companies and how to handle regulatory factors.
- ROI will be a key consideration as analytics projects are evaluated in a post-COVID, lower growth environment.