Data capture solutions: Traditional OCR vs. Cognitive

In the past, all invoices were processed manually; there were data entry clerks in all finance and Accounts Payable departments. This process was slow, fault-prone and not scalable. Employees would spend hours manually rewriting information from paper invoices. Then, along came a solution that used OCR (Optical Character Recognition) technology to extract data from documents when templates and rules were set up. Once the rules were in place, the data would be extracted quickly and accurately. However, users later realized that each new document alteration needed an additional set of rules and templates. That lead to higher associated costs and the AP department‘s dependency on the developer who set up all those rules.

Today:

Manual data capture

Still used in 90% of cases, manual retyping data from documents is painfully slow, fault-prone and not scalable. Accounts Payable employees spend hours manually retyping information which leads to diminished employee satisfaction with the highly repetitive work. Time, money, and resources are being wasted on manual data entry.

Template-based data capture

When it comes to template-based data capture, OCR is the king. Optical Character Recognition is a technology that recognizes text from a document or image. Vast improvements came when OCR was developed, and data did not need to be rewritten manually. Most people thought that this solution solved manual data entry required for capturing data from documents because the output is highly accurate when it comes to documents with low variability.

Companies with hundreds or more vendors have difficulty scaling. With traditional OCR solutions, new templates need to be set up for each vendor and every alteration. This leads to inaccuracies because without the rules, data will not be extracted correctly. It is very time-consuming and expensive to implement new templates. It takes several hours to set up a new template and another 3 minutes for an operator to process a single invoice. Costs for operation include the setup of templates, on-premise implementation, maintenance, and invoice processing fees.

Cognitive data capture

There is a solution that is accurate, fast, and cost-efficient. Unlike a human, Artificial Intelligence has no distractions and needs no breaks. Cognitive data capture can achieve up to 98% accuracy, processing 6x faster than manual processing, with pricing as low as $0.05 per invoice.

The process is very simple: documents are sent (by email, robot, or API) and received into the system. The AI processes the documents, when, if necessary, a human can review, and then the data is exported to be integrated into the accounting or ERP system of your choosing.

Tomorrow:

The situation within data capture cannot improve by itself if we do not change it. We are seeing document volumes increasing, a growing number of layouts with the addition of new suppliers, human operators becoming more expensive and unwilling to work on document transcription, and a rise in security concerns.

The future lies in AI-powered data capture.

It is possible to leave behind the issues of manual and template-based data capture and achieve a solution with high accuracy, speed, and that is cost-efficient.

Benefits to cognitive data capture include:

• Fast Deployment: can be deployed in Accounts Payable process within a few days for any type of invoice or document.
• Effort Reduction: Achieve a 97% reduction in keystrokes to capture the same data and up to 6 times faster compared to manual data entry.
• Continuously Improving: Learn from each invoice processed. Automatically, behind the scenes. 
• Extensibility: Integrate in days through email, RPA or API. Validate extracted data against your ERP. Adaptable to any business environment.

The Future of AI in Data Capture

While traditional OCR solutions may still be a better option for companies with a low volume of documents and few vendors. Setting up templates is more efficient for those companies with little variation to their documents. While there is still an extensive set up process, once the templates are in place, they will capture the data quickly and accurately for documents with the same layout. 

However, compared to template-based OCR, cognitive data capture is able to increase productivity by reducing the number of keystrokes required to perform the same task. Overall, the AI-powered data capture can help you improve time management, accuracy and productivity and is a perfect solution for companies with high document variability. Time management, accuracy, productivity and process improvement, all work together towards increasing your efficiency, and subsequently your bottom line.