Sometimes it’s not all about adopting AI to make things work better. I was reminded recently about the power and significant benefits of process reengineering in a call with a prospective client. Most companies that deal with document-based data within key processes naturally gravitate to the idea of applying AI to automate a lot of those manual tasks. And why not? We, vendors, promise to solve all sorts of problems from processing forms with handwriting all the way to finding proverbial needles in the haystack of unstructured documents like contracts.
But a focus (or sometimes a fixation) on intelligent document processing can lead to some significant missed opportunities. In the case of one client dealing with medical information, a core part of their business is the intake of large patient record files, many of which are based upon unstructured progress reports and other complex diagnostic information. They were seeking a solution that could automatically sort and identify individual documents, remove others, and then perform data location and extraction of key data. With modern IDP software, all of this is within the realm of possibility provided there is an ability to train the system on enough data. But hidden within the main requirements they sought to satisfy were some significant process inefficiencies. Oftentimes inefficiencies are created through the use of software that just isn’t up to the task. In the case of this client, the work they performed every day in their current application was possible but very inefficient. The same tasks had to be done on each individual document instead of in bulk, and oftentimes each task required going through multiple steps instead of just one.
Here’s the kicker: all of this additional work created through these “software-induced” inefficiencies added up to increasing the time required by 200%. Two-hundred-percent! The answer? We didn’t start with IDP at all. Rather, we focused on creating a simplified process that also has a nice side effect of reducing potential errors. Automation of the tasks themselves can save another 50% of the time but it was clear that implementation of automated document ID and data extraction should be a second step, not a first.
In another case, a prospective client, seeking to replace their existing IDP solution with one providing better recognition admitted that they benefited greatly by moving from a basic Excel spreadsheet-based process to a workflow that shepherded staff through each document – and without being successful with their original goal of using automated data extraction!
We’ve seen this time and again: companies seeking to be more efficient routinely step over the possibilities of significant gains in order to implement document automation. Increasingly, IDP vendors are embracing these opportunities to solve the process problem entangling the document problem and it is leading to some pretty exciting solutions. We’re calling it “assistive automation”, which blends the “lights out” automation capabilities of IDP software with the efficiencies of a redesigned workflow.
The upshot: before you zero in on a technology solution, take another look at your process and where the hidden costs are beyond just document sorting and data entry. You’ll often find document automation is only a part of the solution.