Intelligent Automation

Insight from a managing director

Why choose intelligent automation?
Insurance companies want to reduce the time spent processing investigation reports.

Meet Mike…
Mike is the managing director of a major European insurance company who wants to reduce the workload of his claims teams.

The claim backlog
Claims teams at Mike’s organisation spend a lot of time reading and processing large amounts of information in investigative reports. All the work Mike’s team performs happens under time pressure. Mike needed to speed up this process while maintaining quality.

Currently, claims teams receive a massive number of investigative reports. These reports are the first step to establishing whether a claim is valid or not and the financial implications for the claimant. The average investigative report is 13 pages long. Reading and processing the information takes an hour on average.

Human error
Claim officers at Mike’s organisation find themselves processing many claims each day. This volume of work naturally produces a certain level of human error and variation which affects outcomes for claimants. It’s not uncommon for two colleagues to have divergent interpretations of the same report. Mike wants to standardise the process to increase operational efficiency.

AI-driven OCR
Partnering with his CTO, Mike discovered that the best way to increase operational efficiency was to implement an AI-driven OCR (optical character recognition) solution. The aim is to reduce the time required to process investigative reports and increase the pace of decision making without sacrificing accuracy.

To standardise processing, Mike’s team embedded the AI-driven OCR within an RPA solution. This creates a robot worker with intelligent eyes. After a successful proof of concept, the robot reader learns to recognise the various templates the organisation uses. After this point, the bot takes on more complex tasks.

Quicker claims resolution
The robot reader saves Mike’s employees from reading investigative reports and manually identifying pertinent information. The robot reader scans each document and automatically extracts the relevant data. This information then becomes available to customer-facing employees far quicker than before.

This cut the time to process a claim from 62 to 11 minutes which is an 82.26% improvement. This efficiency gain has freed up 23 employees in the claims processing department who he reassigns to higher-value business tasks.

An overview of IA application scenarios

Robotic Process Automation Intelligent Process Automation Augmented Intelligent Process Automation Autonomous Agents
Tuned with structured data Prescriptive Analytics Decision engines with deductive analytics Self-learning decision engines
Only applicable to low-level processing Decision engines Reasoning and fast judgement Exception handling capability
Rules-based decisions Tuned with historical data in structured or unstructured format Knowledge engineering Zero or minimal human intervention
Automatic initiation Periodic pruning and retaining of decision logic Learns from human decisions Superior to human decision
Limited human supervision Limited human intervention for assurance

What's Your Story?


71-75 Shelton Street, Covent Garden
London, WC2H 9JQ
United Kingdom