By: Tom White
The advent of artificial intelligence (AI) has undoubtedly caused a significant upheaval in numerous industries worldwide and is theorized to be the main driver of the upcoming Fifth Industrial Revolution. The International Monetary Fund estimates AI to affect almost 40% of jobs worldwide, including even some high-skilled jobs. As a result, many companies are preparing for a massive shift in their industry landscape, with efforts such as retraining their workforce or contributing to other areas of the value chain, where human ingenuity and decision-making can offer a better result than machine algorithms.
Experts have identified data entry and related fields as among those that will be significantly affected by the proliferation of AI, and many organizations are transitioning from manual data entry operations toward the use of AI technology, such as intelligent character recognition (ICR) and optical character recognition (OCR). However, AI is not a magical solution that will solve all organizations’ data entry and extraction needs.
According to Alan Bandell, Founder and CEO of Axion Data Services, one of the leading US-based data entry services providers, both AI-led and human-led data operations have strengths and weaknesses, and it depends on the project whether one or the other will be better. He adds that companies should exercise due diligence and consult an expert before choosing which method to apply.
Bandell says that using AI requires a high volume of documents because AI processors need to be trained on many documents to accurately recognize and predict patterns, especially when records are handwritten. He estimates this number to be around 500,000 to one million documents a year at a minimum. Naturally, this rules out smaller organizations, and Bandell predicts that a market for manual data entry will remain for some time because some organizations or projects have too small of a data set for AI to function correctly at this point.
Because of AI, Bandell observed that the cost of processing each document went from $4 to $0.20 in one case. He says: “Furthermore, it increases volume processing capabilities and speed of delivery to the client. Industries such as healthcare, insurance, banking, and others that deal with large numbers of consumers are poised to benefit hugely from AI. This also includes the military and government agencies.”
Established in 1996, Axion Data Services has witnessed how much the rapid advancement of technology has influenced the fields of data entry, data extraction, and data verification. Under Bandell’s leadership, the company has adapted to the ever-changing technological and market conditions. One of its guiding philosophies is the Japanese concept of kaizen, which means continuous improvement. Bandell learned about this during his military service when he was stationed in Japan, and he applied it to his business, allowing it to evolve alongside changing times.
Through an industry partnership with an expert in AI technology, Axion can provide comprehensive data services that combine both the speed and high volume capacity of AI and the accuracy and ingenuity of humans. Previously, Axion developed a two-pass process with separate teams of operators, allowing each team to catch any discrepancies or errors committed by the other. Adapting to a new AI-powered workflow, Axion plays on the strengths of humans and AI, allowing them to complement each other.
“However, AI models need significant resources and time to be trained for particular tasks. There are applications where human data entry is no more costly or even less costly than trying to implement AI,” Bandell says. “It’s a matter of volume – if the volume is high, AI is the way to go, but the manual is more economical if the volume is low. But, many projects fall within a wide middle ground, a tossup. That’s when experts are needed to determine what type of data entry and extraction needs to be used.”
Bandell explains that while AI has the advantage in speed, humans have the advantage in accuracy, especially if there is an effective system, such as Axion’s two-team approach. He says that using two separate teams has resulted in an accuracy of 99% – higher than what some clients require of AI, which is usually around 95%.
“No two projects are identical, each with unique requirements and client priorities,” he concludes. “Axion Data Services’ long-standing industry experience allows us to determine the best data solutions for our clients, enabling both efficient and effective results.”
Disclaimer: The information presented in this article is for informational purposes only and does not constitute professional, legal, or technical advice. Readers are encouraged to consult with qualified professionals or experts for advice tailored to their specific needs. The opinions and insights expressed by Alan Bandell and Axion Data Services reflect their perspectives and may not apply universally. Axion Data Services is not affiliated with Unsplash, and the use of their image is solely for illustrative purposes.
Published by Iris S.
