Industry analysts rank artificial intelligence and machine learning as this year’s top technology trend and they promise to rank in the top trends for 2018. For companies with high volume document processing needs that are either processing in-house or outsourcing these tasks, it’s worth examining the level of automation currently used and if AI and machine learning are even under consideration, because this will have a direct impact on efficiency, cost and data quality.
In fact, when 50 top business process service providers were recently asked whether they were considering using AI for data capture, 64 percent said, “yes,” but a remarkable number 36 percent said “no.”
Today, you might be getting away with a heavily manual data entry process, but it’s unlikely you can continue to compete without automated operations in the near future. Human intervention is costly, error-prone and often unnecessary for well-documented workflows. As workers, we’re much better suited to exception handling and higher-level decision-making than repetitive tasks.
The interesting aspect of document capture or recognition is that we think primarily in terms of automating data location and extraction, but automation can and should go much deeper than that. Ongoing measurement and tuning of the capture system to maintain data quality and accuracy are important. Through automated micro-adjustments, the system continues to classify, locate, extract and verify data so the results of document processing provide business-ready data continuously over time.
In Dynamic Production Environments
We all recognize that one-time automation—without business rule updates (either by SMEs or with the aid of AI) of any kind—in our dynamic production environments leads to data degradation and/or intensive manual oversight since documents and images continually transform and additional document/image types get added to the capture system.
Spending millions of dollars manually verifying data is fine as long as everyone is doing it, but as this changes, as Operations become optimized through document processing automation at every level, this spend not only no longer makes sense, it’s dangerous in our competitive market.
If you found this article interesting, you may find the latest 100 Leaders in Capture AIIM survey results of interest: