AI in Insurance Hype or reality – PwC report

Article Synopsis :

Artificial Intelligence (AI) is being touted as one of the most promising technology advancements in human history. As such it’s crucial insurers understand AI and its potential impact on Business as Usual.

Ultra-low operating costs combined with radical innovation – AI holds the potential for real competitive advantage

A new PwC report “AI in Insurance, Hype or reality?” provides useful examples of AI in the insurance industry with real use-cases around customer experience, distribution, risk management, and operations.

AI, as discussed in the report, includes the following capabilities:

• Natural language processing
• Deep QA systems (the use of deep question answering techniques)
• Machine learning (decision tree analysis and Bayesian learning)
• Agent-based simulation modelling
• Virtual personal assistants
• Soft robotics (process mining to identify bottlenecks, inefficiencies)
• Sensors/Internet of Things

According to the report, as AI evolves it will impact the following areas of the insurance value chain:

• Personalized customer experience: Redefining value propositions
• Digital advice: Redefining distribution
• Automated & augmented underwriting: Enhancing efficiencies
• Robo-claims adjuster: Reducing claims processing time and costs

Although AI’s initial impact relates primarily to improving efficiencies and automating existing customer-facing, underwriting and claims processes, over time, the authors believe, AI’s impact will be more profound. AI applications will ultimately identify, assess, and underwrite emerging risks along with formulating new revenue sources.

The report suggests the following specific steps for incorporating AI techniques within a broader data science group, which already exists as a department within most larger insurers:

1. Start from business decisions
2. Identify appropriate AI areas
3. Think big, start small
4. Build training data sets
5. Pilot with parallel runs
6. Scale & manage change

Link to Full Article:: click here

Link to Source:: click here