"$600 million in U.S. shale infrastructure + 2,000 million well-feet drilled = 100s petabytes of relevant data."

Digital Oilfield Outlook Report, GE, Accenture & JuneWarren-Nickles Energy Group, October 2015


Automatically Pull Out Essential Data 

Actionable Intelligence Management & Extraction

Data extraction is a critical part of business process management and should be automatically extracted from forms, electronic documents and communications. It's not possible to have humans read and understand every bit of data coming into an organization. To keep up with this accelerating inbound information flood, extracted data can feed a workflow or populate a database.

Information extraction occurs after incoming communications (web, documents, forms, email) are captured and classified. The data mining process automatically extracts information from structured and unstructured machine-readable documents. Paper documents are scanned and converted to electronic documents via the use of built-in Optical Character Recognition (OCR). Key words (e.g. account number) and identification features can be pulled-out and reconciled with CRM and ERP systems. 

AIM Lifecycle: Document Capture & Analyze

    Extracting data from the plethora of documents in today's enterprises can be a challenge but is an essential part of the Document Capture and Analyze stage of the Actionable Intelligence Lifecycle. This step uses rules to extract and index the relevant pieces of data from the documents.

    Learn more about AI Foundry's end-to-end Actionable Intelligence Lifecycle

    Click to learn more about AIM Lifecycle


    Data Extraction


    Extracting information from unstructured documents, like the one shown to the left, is difficult. Our Actionable Intelligence technology takes a layered approach to automate the process. A rules-based method is augmented with user-guided machine-learning to obtain the highest accuracy when extracting data from unstructured documents. 

    In addition to keying off words, extraction zones can be identified and flagged for manual review if not filled-in.