Most people think that Intelligent Document Processing (IDP) systems are used by companies to primarily save costs. Our experience with some of our leading financial services customers has taught us that this is not usually true. The number one problem that most companies try to solve with IDP is scalability. How can they enable their teams to handle a higher volume of data more rapidly? How can they offer better response times or reduce SLAs for their key customers? While most of the heavy lifting of processing documents is done by the IDP system, you still need a data processing team to handle exceptions. As you scale your business, you often need to add more people to the team.
How long does it take for a new member of your data processing team to be efficient? For some of our customers, that number is 3 months, for others with complex documents with a huge number of variations, that number is closer to 9 to 12 months. Yes, it can sometimes take 9 months for you to stop looking over someone’s work and be assured that they are extracting the right information.
While they are training, how do you make sure that the new members of your team are not making mistakes? The ability to audit data accuracy becomes critical.
Data Correction Audit
There are three main reasons that you will need to audit the corrections made by users.
- You have a new member on the team who is coming up to speed. You would want to audit this person’s work regularly — perhaps even multiple times during the day. The purpose of the audit, in this case, would be to provide feedback and use it as a means for training new users.
- The second and more frequent use of audit functionality is for QA to make sure that everyone on the data processing team is doing their job right. Most of our customers audit all of their users at the same frequency to begin with, but over time, increase audit intervals for people who usually do not have processing errors.
- A third case shows up when you have an obligation to your customers to audit data periodically and produce reports for compliance purposes.
In the past, these audits were done completely manually using physical pieces of paper. But manual auditing is expensive, time-consuming, and resource-intensive.
Infrrd’s Data Correction Auditing
Infrrd’s IDP platform comes with a built-in audit feature for validating corrected data. It is real-time, cost-effective, and can be easily managed by your in-house team. You can start by defining your audit window with things like date range, type of document, users, and STP vs. corrected documents.
This brings up the list of all the candidates for auditing.
Next, you can select the document to audit and see the changes made by the data processor:
You can also record your comments about why a change was incorrect or what should be kept in mind for these kinds of changes in the future:
Finally, you can approve or unapprove the audit data based on your findings, and then export the audit results as a PDF report to share them with relevant stakeholders.
Audit as a Tool For Scaling Operations
You might have heard this saying before — Why do cars have brakes? The answer is a little counterintuitive — to make them go faster. When you scale your data processing operations rapidly, auditing can work as an important tool to help you stay the course and not compromise on quality. Instead of looking over everyone’s work all the time, you can selectively inspect the work of those who are likely to make more mistakes. Infrrd’s built-in audit feature can also help you:
- Evaluate the operational performance of the correction team and users and share continuous feedback.
- Evaluate the performance of new users to provide relevant training at an early stage.
- Identify and resolve errors immediately on receiving support tickets from customers.
- Maintain a performance record for your team and find people who are not necessarily the fastest, but are the most accurate.
Infrrd’s IDP platform comes out-of-the-box with this powerful audit feature to help you scale your operations.
Originally published at https://www.infrrd.ai.
