RPA & Computer Vision: 5 Intelligent Automation Examples in '24

With hyperautomation becoming one of the top enterprise technology trends, leading robotic process automation (RPA) companies are beginning to integrate artificial intelligence capabilities into their automation tools to create intelligent bots that automate end-to-end processes and make decisions. Computer vision is one of the key AI capabilities of new-generation RPA tools.

In this article, we will explore five examples of how computer vision and RPA can be used together to enable business intelligent automation.

1. Remote desktop automation

50%¹ of businesses expect to increase their spending on remote desktop software as:

• Nearly 60%

However, such virtual desktop infrastructures (VDI) only provide access to images from applications to end users, and traditional RPA and automation methods that require operating system access cannot automate processes through a remote desktop.

Intelligent RPA bots with computer vision capabilities can “see” the screen of a computer as human employees do and:

• Record the user’s interaction with the screen to learn how to carry out a task.
• Interact with user interface elements such as images, texts, buttons, dropdown menus, and checkboxes.
• Open and close applications.
• Extract data from an application and enter it into other applications.

Learn more about RPA and BPO integration.

2. Legacy system integration

Legacy systems are pervasive in many industries because replacing them can be disruptive to essential business operations. For instance:

• 80%

While these systems can be challenging to replace, they can become obstacles to digital transformation in the long run. Intelligent automation can offer a solution for businesses that want to integrate their legacy systems with modern applications without developing specialized integration solutions. Computer vision-enabled bots can:

• Connect to different types of software including legacy and modern cloud applications,
• Interact with GUI elements,
• Extract and migrate data between applications.

3. Handwriting recognition

Figure 1. Cursive handwriting is hard to recognize.

Although we have optical character recognition (OCR) that can recognize text printed in different fonts since the 70s², recognizing handwriting is still a challenge for machines as:

• Handwriting varies greatly among individuals,
• An individual’s handwriting style can even change over time,
• The handwriting might be cursive or skewed, which makes identification difficult (Figure 1).

Neural network-based computer vision models can achieve 99%³ handwriting text recognition accuracy. Leveraging these models, intelligent bots can digitize handwritten documents in industries such as:

• Healthcare: Prescriptions, patient forms
• Financial services: Cheques, KYC documents, money transfer receipts
• Insurance: Claim documents, accident reports

4. Insurance claims automation

Efficient claims processing is important for insurance companies as:

• Claims processing accounts for 70%

Combining RPA and computer vision can improve insurance claims processing. For example, when an insurance company receives a claim for car damage, an intelligent bot can:

• Guide the insured to take photos or videos of the damage using conversational AI,
• Analyze the visual data provided by the customer,
• Generate a report to be used by the insurance company to determine the extent of the damage, decide on the claim, and prevent inflated claims.

Bots with computer vision capabilities can also be used to verify the identity of claimants. This is important in cases where fraud is suspected. Bots can extract information from ID documents and compare images to verify a customer’s identity and help insurance companies avoid paying out fraudulent claims to non-existent customers and entities.

5. Banking customer onboarding

Figure 2. COVID-19’s impact on the use of digital channels in banking. Source: BCG

When it comes to customer onboarding, banks are under pressure to provide a fast, convenient, and compliant experience as:

• The COVID-19 pandemic has accelerated the shift to digital channels for consumers, such as mobile and online banking (Figure 2).
• The average rate of customer attrition due to onboarding issues is between 25%

Banks can leverage intelligent automation to automate the know your customer (KYC) process for faster and compliant customer onboarding. Intelligent bots can:

• Capture data from documents such as ID cards, driver’s licenses, or utility bills using OCR. This data can then be automatically entered into the bank’s customer information system, reducing the need for manual data entry.
• Verify the identity of applicants by comparing their facial features to those on ID documents. Facial recognition can prevent fraud and improve the accuracy of KYC checks.

If you want to get started with cognitive automation, you can check our in-depth article on intelligent automation tools. If you have other questions, feel free to ask:

External Links

• 1. “Remote Desktop Software Statistics and Trends“. (2020, May 26). TrustRadius. Retrieved November 10, 2022.
• 2. “Optical character recognition“. (n.d.). Wikipedia. Retrieved November 10, 2022.
• 3. Wiles, R. (2019, June 27). “Have we solved the problem of handwriting recognition?” Towards Data Science. Retrieved November 10, 2022.