According to Gartner, “The market for document capture, extraction, and processing is highly fragmented. Data and analytics leaders should use this research to understand the process flow and differentiated capabilities offered by intelligent document processing solutions”. Gartner’s recently released “Infographic: Understand Intelligent Document Processing” covers these 6 critical flows in IDP.
1. Capture or Ingestion
2. Document Preprocessing
3. Document Classification
4. Data Extraction
5. Validation and Feedback Loop
This is the fifth and final post in the series where we explore Integration. Check out our earlier posts in this series, Capture and Preprocessing, Document Classification, Data Extraction, and Validation and Feedback Loop.
Meaningful data offers the best benefits when they are integrated with your business or enterprise systems, be it your on-premise or cloud system, or any incredibly complex system, such as an ERP. Today, businesses are focused on formulating comprehensive solutions for constantly-evolving customer problems or needs, and it is important to have an integrated system to ensure greater efficiency and business effectiveness.
When it comes to Business Intelligence (BI) & Analytics, unstructured data has been kept outside of data mining for the longest time. If you run a retail clothing store, when you sell a dress, you record its sale, you capture details like selling price, payment method, discount, tax, etc but you do not record how the dress looked. Did it have half sleeves or full, what kind of neck design it had. All of this information is potentially in the photo of the dress. This limits you from understanding your customer behavior. Questions like what percentage of people who buy faded blue jeans pair it with belts featuring over-sized buckles.
In the absence of a system that can make sense of unstructured data, it was always kept outside the realm of BI and Analytics. Structured data, like your sales record, also happens to be a small fraction of the overall data that you have access to. The majority of data that any organization deals with is unstructured data such as emails, documents, receipts, and photos. Now that IDP platforms can convert this unstructured data into structured data, it opens up exciting new avenues of understanding your customers and their behavior better through data mining.
Here are a few examples:
- From a receipt of other stores that you do not own, you can now figure out if people who buy a beer also buy wine. If you find they do, you could run a promotion selling them together.
- From payslips in mortgage application documents, you can figure out that most people who work for sales in the manufacturing industry usually get only X% of their sales commissions.
- From supporting insurance claim documents, you can automatically figure out what percent of a car repair cost is from body shop work vs replacement parts for a Toyota Prius serviced in Chicago.
You can take this analysis one step further by opening up your extracted data to search using Natural Language Query (NLQ) technologies. So, instead of setting up reports in advance, you can fire a query in natural language. If we had an automated assistant, you could ask, “How many mortgage applications did we receive for homes in the bay area yesterday?” And you would get the right answer.
Some of the common features to check out in an IDP platform to evaluate their integration capabilities are as follows:
Some of the common methods used for IDP integration with third-party solutions are as follows: