PandaTS Collaborates with Google Cloud, Bringing AI to Document Verification

Surging interest in online trading during recent lockdowns has led to new account registrations going through the roof.

Tedious document verification procedures have long been the bane of back office staff. The process is repetitive, time-consuming, and can be a troublesome bottleneck for brokerages during periods of peak traffic.

Surging interest in online trading during recent lockdowns has led to new account registrations going through the roof. This has inundated back office and compliance teams with KYC/AML documents to inspect and verify.

“The sheer amount of documents to process can temporarily cripple certain brokerage departments,” PandaTS CEO, Samuel Gutman explained. “It can be a mechanical task at the best of times, but when new traders are flooding in, it becomes more complex,”

Speeding up and optimising the process would undoubtedly ease the burden for everyone, as brokers could avoid tying up large numbers of employees, and traders wouldn’t have to wait as long to get verified during peak times.

“In our experience,” Gutman continued, “it’s often the CRM that causes some of the friction when trying to verify large numbers of documents. Whether it’s the process of pulling them up, or navigating back and forth through menus, there’s usually something there to frustrate the person who has to repeatedly perform the task. At PandaTS, we’ve tried to make this process as efficient and intuitive as possible in our own Panda CRM, especially with the new update that makes finding, reviewing, and comparing KYC documents against registration info much easier and smoother. But when each member of staff can only process one application at a time, you’re still going to get bottlenecks during busy periods.”

Enter Google Vision AI

In collaboration with Google Cloud, PandaTS recently launched a new module for Panda CRM that harnesses the powerful image detection AI of Google Vision and applies it to the task of document verification.

A common reason for documents being rejected is that scans and photographs provided by would-be traders are often either of bad quality or have been improperly cropped.

Vision AI

Ordinarily, each failed verification request has to be handled by a human member of staff. This only creates more work downstream as the documents have to be resubmitted and reviewed.

However, thanks to Google Vision AI, the new document verification module from PandaTS allows documents to be automatically checked for quality and rejected

As a first line of defence, the module can handle much of the initial hassle of document verification on its own, freeing up human staff for more pressing and value-adding tasks.

But, as Google’s Vision AI team explained to us, it can do more than that. Vision AI’s sophisticated optical character recognition capabilities allow it to actually detects and understands text in the supplied documents.

So, whenever a new document is scanned and uploaded by a client, Panda’s verification module can review it for quality, but also check it against the information the client has already provided during the registration process.

It compares the information from the scanned document to the customer details stored in the Panda CRM, then assigns each document a score depending on how accurately it matches the details already held on file for that client.

PandaTS is expecting the new module to be a game changer for its brokerage clients. Helping back office and compliance teams by filtering out bad images without tying up any members of staff, but also allowing them to segment different uploads, quickly approving high-scoring submissions, and batching the lower-scoring ones to be delegated to appropriate team members.

Gutman anticipates that the module’s abilities to read, compare and score incoming documents will reinvent how online brokerages handle their KYC/AML commitments, as well as providing them with a competitive advantage during busy periods.

“To start with, we expect the new module to eliminate uploads that have to be manually rejected for poor quality. After that, the scoring system streamlines the process even further by allowing perfect matches and high-scoring applications to be handled differently than the low-scoring ones. It’s like having a faster member of staff doing all the tasks you hate,” he joked, “Not only that, it can handle more than one verification at once, and doesn’t even need to take breaks! Furthermore, with machine learning, our document verification module will actually get smarter as it’s progressively trained on more data. In other words, its accuracy will increase the more you use it.”

About PandaTS

Panda Trading Systems specialises in the development of superior brokerage technologies for the retail trading industry. The company’s mission is to make online trading a smoother, more integrated experience, for traders and brokers alike.

With 15 years of industry expertise under its belt, today Panda serves many of the world’s leading brands with its suite of customisable brokerage solutions, allowing them to optimise their operations and freeing them up to focus their resources on growth.

Got a news tip? Let Us Know