At the end of 2017, following the acquisition of Scottrade, TD Ameritrade faced a massive task: extracting and reconciling co-applicant account information across over one million pages in order to comply with KYC [Know Your Customer] regulations. The company had one year to extract and reconcile the information into their new system, so they began seeking a solution to help them do it.
The forms they received included a mix of handwritten text, faxes, and low-resolution images that legacy data extraction technology like OCR was unable to read and process. Before discovering Hyperscience, TD Ameritrade estimated it would take 60 full-time keyers looking at the images and keying in information (e.g. name, DOB, SSN) to complete the project and meet the deadline.
The Hyperscience Platform’s ability to automate without sacrificing accuracy was paramount, especially for an organization that deals with sensitive customer data where every digit needs to be correct.
Hyperscience was able to deploy the technology quickly, getting it in place, and then letting the machine learn and improve over time which fitted with the company’s agile mindset.
By automating manual tasks, TD Ameritrade has shifted the potential of its workforce, allowing for less data entry and moving those roles to more knowledge work and process engineers. The shift has allowed the company to focus on higher value work, such as analyzing data, improving existing workflows and making efficiency gains. With Hyperscience, TD Ameritrade was able to double processing speed with 50% fewer data keyers, quadrupling throughput within the first 4 weeks.
“The ability to get a 90 plus percent transcription rate was impressive. We certainly weren’t getting those results from some of the other folks that we were contacting, which were quoting us closer to the 50-60% range.”