There has recently been a lot of chatter in our industry about machine, and I want to share a few thoughts about how ImageSource’s ILINX Advanced Capture 8.5 platform tackles this topic. Our approach is to deploy a neural network-based document processing model that does not rely on templates. Our machine-learning platform supports custom-developed content classification projects with much faster turnaround than traditional rules-based models. The result is significantly faster time-to-production with more reliable and accurate results for our customer partners. ILINX machine learning offers:
Our solutions leverage machine learning to create pre-built form classification algorithms in the lab, which provides a more-flexible and efficient way to develop new document processes.
For the sake of this article, I’d like to focus on the goals of almost every AP automation project:
For the initial document discovery, a technique called “clustering” can be used to automate the logical grouping of like documents. Clustering, in this example, refers to different categories of like invoices, checks, receipts and remittances. Documents can be organized automatically. Invoices from one vendor can be grouped together, such as receipts to travel documents. The result is a set of documents grouped by likeness that can then be further evaluated.
Next, each cluster, if part of a required document, can be given a document type (or class). The training set can then be imported into the machine learning ILINX solution designed to automatically identify key characteristics of each document type (often called “feature extraction”). This trains the neural network for each document type. When performance is not ideal for a specific class, the customer can add those misclassified or unclassified documents to the class sample set to “re-train” the neural network.
Data extraction is simplified by taking sample invoices that have been processed, along with the data required for each document. Together, these automatically train the software to locate the matching data and derive positional algorithms for each data field. The software uses the processed data for each page and locates the corresponding data on every document. The solution will do this for each sample and then automatically create algorithms based upon exact location, changes in placement across each example and relative position to other data, among other elements. The knowledge worker simply examines the results.
The technology used to configure the system also makes real-time adjustments. Complicated projects that typically would take weeks, if not months, are significantly reduced. Machine learning technology streamlines the manual processes used in production and helps reduce overall labor costs.
This effort can be applied to automate both paper-based and electronic document-based processes in a single workflow.
by Terry Sutherland, CEO, ImageSource
If you would like to learn more about how ILINX machine learning can automate your business please contact us at firstname.lastname@example.org or
Forrester Research analyzed and evaluated ECM technologies and came out with The ROI of Imaging. Forrester Research, Inc. is a global leader in business and technology. They define imaging as software for scanning, capturing, indexing, retrieving, processing and archiving digital images of documents and electronic forms. Many organizations rely on paper intensive business processes and because of that, imaging is a very important component of Enterprise Content Management’s value.
Great support is a must for any business and Customer Service 101. Automated support is great if you want to pay a bill with a credit card or check the balance of an account, but when you have an actual issue you need assistance with, automated support is the LAST thing that I know I want. You call the number, have to start pressing buttons and then keep getting dumped into various queues, and then when you actually do reach someone, sometimes you explain your issue, and then you get transferred 10 times until you reach the right person. I honestly dread calling any company or organization where I know this happens, because I know it’s not just a quick 5-10 minute call. It always ends up being 30 minutes to an hour of my time at least!
That’s one of the things that I think is so awesome about the Support Team here at ImageSource. Our customer partners can put in a support ticket via the website with information about the issue they are having and a live human being from our Support Team will call them to help work through the issue. The team can chat with the customer over the phone or even set up a WebEx session to dial-in and see what’s going on. What’s even more wonderful about working with so many great organizations in the Olympia area, if the issue can’t be resolved over the phone and/or over a WebEx session, a technician can be onsite relatively shortly to assist. How cool is that?!
On top of all that, there is always someone from our Support Team on call 24/7. So if a customer partner has an issue in the middle of the night that needs immediate attention, someone is always on call if the need arises!
ImageSource’s Support Team is very knowledgeable on a number of different content management products, from capture software to eForms to records management and everything in between.
So if you’re having an issue, don’t be afraid to reach out. You’ll get to talk to a human and you won’t be put on hold for an hour…
We recently helped the largest credit union in Alaska (17th in the nation) automate their consumer loan processing operations. With a growing number of their 77 locations looking to processing auto dealership loan requests, Alaska USA Federal Credit Union needed to replace their manual, paper-based process to meet current and future demands.
As is the case with many lenders who haven’t yet gone electronic, paper loan requests packets were being submitted via multiple avenues from dealerships. These packets consisted of anywhere from 20 to 25 pages per packet with a variety of page sizes, file sizes and quality. Some documents were even printed from DOT Matrix printers at some of the dealerships. Packets came in primarily through fax and email, and local dealers in Anchorage would even hand deliver loan packets. Once received, faxes and emails were printed out and paper documents were used to process the loan request. The documents had to go through various people for review and approval creating a lot of manual movement of paper. Having these documents in paper format also created a need for document storage in file cabinets consuming significant floor space in the building.
In order to rise above the limitations that paper processing placed on loan approval volume, Alaska USA went digital, with 6 major automation improvements to shave valuable time off the process:
Having automated their Consumer Loan process, Alaska USA has greatly improved their service to auto dealers and can be more competitive on winning new business. On the cost savings side, they can process more loans without increasing headcount and have eliminated hard copy document storage.
Migrations from systems like IBPM to ILINX can be fraught with issues that can bite the unwary in very bad places. However, if you are aware of such problems, you can plan ways to mitigate them and have a successful migration in the end.
One issue we run into is documents that have a page or two with corrupt images. Perhaps when the page was first contributed to IBPM, a system or other type of issue caused the image to be corrupt or cease to exist. Either physical hardware or a software bug can be the culprit. The product we use for migration, ILINX Export, will flag this document as an error, skip it and move on to the next document in RECID order. Once the export is completed, these flagged documents have to be re-visited. Once a determination is made that an image is indeed corrupt, and the chance to recover it from backups is extremely remote, the document can be deleted or manually exported from IBPM without the corrupt image.
Another matter we’ve dealt with is related to non-tiff images. This category is “universal” type images, and includes PDF, DOC, XLS, MSG and a host of other file types that IBPM supports. There are options within the ILINX Export tool that will allow the export of these files types in their native format through the IBPM SDK. Or the export can be done through database manipulation that can directly access the image file and then “unzip” the universal file into its native format.
The issue that can be encountered here is twofold, and manifests itself when migrating to another repository. One, IBPM stores the native file zipped up with another file that contains metadata and has no file extension. When the document is unzipped there are two files, one with a valid file type and one without. Typically, backend repositories require file extensions, which are useful for performance, like displaying file type icon on the user interface, and a variety of other reasons. During the migration, importing to the backend may be impeded due to a lack of extensions on the metadata files. Secondly, if the extension of the universal file has been altered or damaged in storage, the file type may not be a standard that the new repository will accept. In any case, having your migration come to a screeching halt is something to avoid.
Awareness is the key. By proactively incorporating a response into your migration plan, you can eliminate much heartburn and anxiety. That is where the expertise and knowledge of a seasoned Optika / Stellent / Oracle integrator, like ImageSource, comes into play. We have helped many customers build migration plans that take these and other items into account, so the migrations are as smooth and worry-free as possible.
Sr. Systems Engineer