New research from data leader, Qlik reveals how concerns around trust and regulatory compliance, as well as weaknesses in the data pipeline, are slowing the adoption of predictive analytics in Financial Services (FS), preventing insurance organizations from reaping the full value from their data.
The report “Unleashing the Potential of Predictive Analytics in Financial Services”, which surveyed 503 UK IT leaders in financial services organizations, of which more than 200 held roles in insurance, exposes the slow uptake of predictive analytics. Two-thirds (65%) of British insurance institutions have five or fewer predictive analytics use cases in implementation – in stark contrast with the leading 9% of institutions which have each introduced 40 or over.
Key issues IT leaders in insurance face when implementing predictive analytics include:
- A question of trust – Every decision a financial services organization makes can have a major impact on a customer’s life, from agreeing to an overdraft to making payday or approving a mortgage application. Yet one third (29%) of IT leaders in insurance admit fearing algorithms could unfairly impact their customers. This is perhaps unsurprising given only half of respondents (50%) trust decisions made by predictive analytics solutions are without bias and are always accurate (47%). In fact, only half of those working in insurance (50%) trust predictive analytics to always give them the best price on their car insurance.
- Regulatory risk – In such a highly regulated industry, 39% of IT leaders in insurance fear they could be held personally responsible for decisions automatically triggered by predictive analytics solutions. The regulatory burden also weighs on them, with 37% reporting it outweighs the benefit the solution could offer.
- Flaws in the data pipeline – IT leaders in insurance also cite a number of technical barriers to implementation. Two fifths face issues with data quality (44%), data silos (41%) and the speed of data integration (36%). Data privacy (35%) and the use of inaccurate or outdated data sets (35%) were also common concerns. Almost half (45%) also fear they don’t have the skills to implement predictive analytics.
Improve trust by marrying human & machine intelligence
Many of the concerns relating to predictive analytics are underpinned by a lack of human oversight of its decisions – both in terms of outcomes and explainability. To overcome these issues, two thirds (65%) of IT leaders in insurance advocate incorporating predictive analytics into business intelligence (BI) platforms.
Most believe integrating the powerful forecasting of predictive analytics into the BI platforms that already inform employee decision-making has the potential to:
- help organizations comply with regulatory frameworks (72%)
- identify areas of cost saving (72%)
- significantly improve employee decision-making (71%)
- deliver a better customer experience (70%)
- democratize forecasting (67%)
However, ensuring employees have the requisite data literacy to understand, question and apply the predictive forecasts to their decision-making process is key to maintaining trust and compliance. Three quarters (77%) of IT leaders in insurance highlighted the importance of data literacy training in enabling employees to recognize the limitations of the technology. And in helping them explain to customers and other stakeholders how decisions using predictive analytics are made (73%).
“The insurance industry is starting to see the value that can come from personalized insights provided at the right time and in the right way. Doing so can provide real value and transform user experience without eroding trust that is a risk when raw data is shared or sold,” said Nick Blewden, Head of Analytics at Lloyd’s of London.
“At Lloyd’s we use the Insights Hub to show each Lloyd’s insurer their portfolio with personalized benchmarks and market intelligence to support their strategic decisions in an easy-to-understand way. Lloyd’s are providing powerful analytics to all insurers in the global Lloyd’s market for free as they realize the power the right personalized analytics can have in transforming business outcomes.
“Better use of data analytics to understand your business, refine your portfolio, remove cost from your operations and plan your future strategy will separate high and low performers in the future.”
“The financial services industry is undergoing rapid data transformation. Predictive analytics will play a key role in empowering employees to take more informed actions, with forecasts helping them consider what might happen, as well as what has happened before, when making decisions,” said Adam Mayer, Senior Manager at Qlik.
“However, our research has shown that many IT leaders are yet to fully trust the insights from predictive analytics and the impact these decisions could have on their customers,” continued Mayer. “This trust needs to be built from the ground up. Real-time, hyper-contextual information, with clear data lineage and robust governance, must feed the analytics data pipeline, revealing insights that data literate employees can discerningly use to inform decisions. This will empower financial services organizations to look forward and take action, rather than react to business moments as they arise. Helping them truly achieve Active Intelligence from their data.”
Find out more in the full report here: Unleashing the potential of predictive analytics in financial services