RESEARCH REPORT

In brief

In brief

  • Automation is here to stay, in insurance and beyond. The lifeblood of virtually every form of automation is data.
  • But there’s a catch: It is estimated that up to 75 percent of an insurer’s data may be inaccessible to automated systems. This is called “dark data.”
  • New approaches to data extraction through artificial intelligence could shine a light on dark data and help automation reach its full potential.
  • Read more on how we can help insurance firms get their automation efforts to the next level by leveraging our insights, tools and industry knowledge.


Businesses of all kinds are no longer just talking about automation or experimenting with it—they’re embedding it into their operations and their long-term strategic objectives. Recent estimates suggest that spending on automation is going to double from today’s levels to reach $97.9 billion in 2023.1 For insurers, looking ahead the big question is how to automate at scale.

The reason behind this investment is clear: Automation promises a quantum leap in process efficiency, especially for information-intensive businesses like insurers.

But there’s an elephant in the room: dark data.

Dark data’s long shadow

“Dark data” is simply the data an organization has but can’t make use of.

Consider a picture of the scene of a car accident. The picture might contain lots of useful data for an insurer—not only the car make, model and license plate number, but also the time of day, the weather and (if the picture contains a street sign) the location of the accident. This information on its own might be enough for an insurer to open a claim and produce an estimate of damages. But without a system to extract all that information, it remains “locked” inside the photo and likely left on the shelf forever.

Dark data is everywhere in insurance, from handwritten documents to voicemails to videos captured on smartphones. And it casts a long shadow.

75%

Up to 75% of a corporation’s data may be dark data2

6-20M

Between six and twenty million dollars is the range of savings for using straight-through processing within a typical insurance claims organization3

Dark data is everywhere in insurance, from handwritten documents to voicemails to videos captured on smartphones. And it casts a long shadow.

The existence of dark data has hamstrung insurance automation by creating the need for “handoffs” between human operators and robots along the automation pipeline. For example, one automation software developer suggests targeting only 70 percent of a process for automation.4

One of the great benefits of automation is helping insurers realize “straight-through processing,” or STP: a reduction in transaction processing time. It was long thought that setting up an automated system that can “ingest” data in a wide variety of formats and then process it seamlessly is an unrealistic goal.

But not anymore.

Use case: STP in insurance

In the aftermath of an accident, a customer uses their insurance carrier’s mobile application to take pictures of the accident. After the pictures are uploaded to the app, a pre-trained model assesses the damage to the vehicle, confirms that the insured is covered, and estimates the cost of the repairs.

The model also suggests the vehicle appears drivable, so the system notifies the customer of nearby auto body shops that the carrier has worked with in the past. Using data from past repairs the shop has done, the system verifies that the final cost matches the estimated cost of the repairs, sends a payment to the insured, and closes the claim.

The power of automation, unleashed

The potential of a model like this for insurance is obvious. But in a data-driven world, the benefits of unlocking dark data stretches far beyond empowering STP. Shining a light on dark data can provide insurers with a competitive edge by feeding the models and analytics that create sales opportunities, measure risk and inform business decisions.

Already, we estimate that insurers who are leaders in analytics can see an increase in profit of between 16 and 21 percentage points.5 As digital change continues to remake the world and create new sources of data, this advantage is likely to grow.

Light up your dark data

In a new report, Accenture explores the consequences of dark data—and lays out a four-step approach for helping insurers unlock its power. Insurers applying Accenture’s framework can not only realize STP and better leverage automation, but also generate benefit streams that pay competitive dividends for years.

1.“Worldwide Spending on Artificial Intelligence Systems Will be Nearly $98 Billion in 2023, According to New IDC Spending Guide,” IDC, September 4, 2019. Access at: More than six hours of our day is spent online.

2.“Dark Data Research Reveals Widespread Complacency in Driving Business results and Career Growth,” Splunk, April 30, 2019. Access at: Dark Data Research Reveals Widespread Complacency in Driving Business Results and Career Growth.

3.Based on Accenture’s experience and work in this area, assuming a claims organization with 6000 FTE.

4.“From pilot to full scale RPA deployment – A comprehensive guide to the business transformation journey,” UiPath. Access at: From pilot to full scale RPA deployment.

5.“Harnessing the data exhaust stream: Changing the way the insurance game is played,” Accenture 2016. Access at: Accenture Insurance Technology Vision 2018.

Duane Block

Managing Director


Geoffrey Hodgson

Manager

MORE ON THIS TOPIC

Neural networks in financial services
Taking the next steps in your neural network AI journey

Subscription Center
Stay in the Know with Our Newsletter Stay in the Know with Our Newsletter