It’s Time to Upgrade Your Data Entry Specialists to AI Associates

Data entry specialists should be extinct by now. At the very least, they should be a rare breed. However, companies are still hiring them to enter data manually into their business systems. Salaries for this role may be low, but ultimately, manual data extraction comes with a high price tag and unnecessary risk. Fortunately, automated data capture solutions enable you to transform costly, error-prone data entry clerks into productive, value-creating AI associates. 

To demonstrate the benefits this transformation can bring to your company, we’ll spend some time with data entry specialist Adam and AI associate Julie.

As you can see in the video below, it takes Adam 8x longer to process the same number of invoices as Julie:

Now, let’s get to know the stars of this clip for better insight into how their data capture methods impact their respective companies.

Adam the Data Entry Specialist

Adam is an accountant working as a data entry specialist.

Adam is an accounting professional working in his employer’s accounts payable (AP). According to his job description, his main responsibilities include:

  • Sorting, reviewing, and confirming invoice payments
  • Entering invoice data into the company’s accounting and enterprise resource planning (ERP) systems
  • Monitoring expenditures and processing expense reports
  • Archiving and maintaining historical records
  • Analysing expenditures
  • Maintaining accounting ledgers
  • Identifying opportunities to reduce costs
  • Writing monthly reports
  • Deliver documentation for audits
  • Providing support for vendor relations
  • Identifying vendor discount opportunities for early payment
  • Identifying payment discrepancies

The reality of Adam’s role, however, has him spending nearly half of his working hours doing manual invoice data entry – one task out of the many listed in his job description. With less time to handle the rest of his responsibilities, Adam often finds himself either working several hours of unpaid overtime or simply letting his additional tasks slide.

So when Adam arrives at his desk, he typically has to start his day processing an assortment of paper and digital invoices. The term “invoice processing” is a fancy way of saying typing or pasting text and numbers into a spreadsheet, accounting system, or ERP solution.

While Adam is fairly efficient, it still takes him nearly two minutes to enter data for each invoice. Once he’s entered the data, Adam needs to double-check his work to make sure he hasn’t entered anything incorrectly. The last thing he wants is a data entry error slipping through the entire invoicing workflow.

Even errors that get flagged before the invoice approval and payment stages come with a price – the process of identifying and correcting errors then verifying corrections adds close to six minutes per invoice. In addition to the extra expense and lost productivity, this could delay payments, potentially straining vendor relationships.

Adam spends his workdays in a perpetual state of boredom and stress. As we’re about to see, there is considerable pressure on him and his comrades in the data entry trenches to be 100% accurate, 100% of the time. Failure to do so could have catastrophic consequences for his team and his employer.

Data Entry Errors in the News

Recently, Adam noticed that data entry errors have been making headlines. Most involve data related to COVID-19 rather than invoice data. In some cases, the mistakes were simply embarrassing; in others, they created complications for economists tracking coronavirus’ effects on the labor market. These cases include:

The High Financial Cost of Faulty Data Extraction

Micheal Ashmore, a Portfolio Director at Precisely, sums up the impact that data entry errors can have on businesses perfectly:

“Dirty data is to a business what dirty drinking water is to a human. It fundamentally affects the vitality and effectiveness of its operations and can ultimately lead to its demise.”

Adam is aware of this, and of the fact that bad data costs US companies and institutions a total of over $3 trillion annually. The last thing he wants to do is cost his employer millions of dollars or cause his company to shut its doors. There are lots of horror stories about data entry errors in business, such as:

Nobody, especially not Adam, wants a mistake of such proportions in their job history. 

Compounding Adam’s underlying stress is a persistent concern about finishing his other accounting duties on time, as well as the boredom and sense of purposelessness that can, at times, be overwhelming.

Fortunately, Adam has the potential to steer his career along a path similar to the one his friend Julie is on. 

Julie the AI Associate

As an AI associate, Julie can do her job as an accountant, without data entry.

Julie’s work as an AI Associate is demanding, and, like any job, it has its stressful moments. Without the relentless tedium of manual data entry, she’s mentally and physically capable of not only managing the demands and stresses of her role, she’s able to reflect on and learn from them.

While her job title may be different, Julie’s got pretty much the same responsibilities as Adam. In the context of finance and accounting, the distinction between what makes a good data entry specialist and what makes a good AI associate is simply this: the former must be a fast and accurate typist, the latter must be a fast and accurate scanner.

Julie spends very little of her workday entering data. A few months ago, her employer implemented a cognitive data capture solution to automatically capture invoice data and export the data to its accounting and ERP systems. 

She was surprised at how the provider deployed its customized solution in just three days – previously Julie had been using a template-based optical character recognition (OCR) solution. The new data capture solution is capable of reading invoices and recognizing the data it needs to capture, so she no longer has to wait for IT to set up new templates and rules every time a new invoice comes her way.

That’s not to say the new solution’s artificial intelligence (AI) does everything for Julie; she still needs to review any captured data that has low confidence scores and make corrections where necessary. However, compared to manual data entry, this takes no time – she now spends an average of around 18 seconds per invoice compared to Adam’s two minutes. All she needs to do is scan the UI, which clearly shows her what data needs to be checked. When Julie needs to adjust a field value or name, she simply clicks and types. Unlike Adam, she doesn’t need to review vast spreadsheets and continuously look from paper to screen or, arguably worse, bounce between PDF and spreadsheet on a single monitor.

Julie’s validation of captured data not only ensures quality data enters the AP workflow, it also helps improve extraction accuracy. In the time Julie’s been working with the AI-powered invoice data capture solution, the software’s precision has gone up to just over 97%.

From Data Entry Specialist to AI Associate in 60 Seconds

Sure, sometimes we’re talking about a few minutes rather than 60 seconds. But the data capture solution Julie uses still processes documents so quickly it takes some people by surprise the first time they experience it. For someone like Adam, it’s life-changing. For your company, it’s game-changing.

So how does it actually work? Julie’s journey with AI started with Rossum’s Generic AI Engine, which could immediately process invoices with astonishing accuracy. She still had to occasionally make corrections, which were minimized once engineers from Rossum’s Customer Success team trained a Dedicated AI Engine to suit Julie’s specific data extraction requirements. This engine also enables Julie to process all her other business documents, such as purchase orders, bills of lading, and any other document with custom data fields.

We’ve helped transform data entry specialists into AI associates, and in every case the results, including efficiency boosts, productivity gains, and cost reductions, have been impressive:

  • Rossum onboarding takes just 10 minutes compared to half a day to learn how to work with a template-based solution, saving managers one month’s worth of training time.
  • Employees can process documents and export their data in less than a minute, giving themselves the time they need to handle their regular duties.
  • One company that adopted Rossum saw a 70% reduction in man-hours spent processing documents.

Adam and Julie may appear to be fictional characters in our Rossum Data Entry Championship video; however, they really exist. The pain of the Adams of the world is real – data entry specialists know all too well the boredom and anxiety that comes with their role. It pays the bills, it’s temporary, and it feels like punishment, especially for people like Adam who have the education and training to do far more for their employers.

The Julies, on the other hand, have the opportunity to advance their careers – as AI associates, they spend little time processing documents so they can actually do their jobs. It would be naive to say that everything is sunshine and roses for all people like Julie, but they can do more for their companies, and enjoy more fulfilling roles in which AI enables them to advance along their chosen career paths.

Leave comment