All Articles

Document Understanding: What it is and how it works

Nov 29, 2021 Digital Transformation Document Intelligence Document Understanding

Document Understanding: What it is and how it works

Click the image to view infographic

Document understanding is the use of OCR, and AI-powered by Machine Learning and Natural Language Processing within a suite of interconnected applications to automatically extract, classify, interpret, contextualize, and search information from documents. Major use cases include reviewing text within legal contracts, leases, and other documents to identify new business opportunities, mitigate hidden risks, and automate business processes.

Document Understanding AI  is able to identify concepts and related information that traditional keyword or “fuzzy” searches would miss. Outputs of an intelligent document understanding solution include useful, actionable metadata about the analyzed documents. For instance, the renewal dates or notable clause wordings contained within them. Additionally, it offers a clear view into each document’s larger context and meaning. Although document understanding solutions vary considerably in their functionality and time-to-value, they generally cover at least the following four major functions:

1. Self-Organizing Document Management

A full-functioning document understanding platform uses its document interpretation engine to process and classify documents, whether they’re uploaded one by one or by the thousands at a time. Documents are classified by type and they are further organized with the additional metadata extracted from them into taxonomies that are far easier to navigate than folder-based systems. Archaic and disorganized folder hierarchies are a thing of the past! Along the way, all this metadata and the full document text are indexed so that it can be searched as quickly and as intelligently as possible later on when a specific document or term needs to be retrieved. AI models continuously learn taxonomy for the processed documents based on user feedback and activity.

2. Contract Analytics

A complete Document Understanding solution combines modern document management with advanced contract analytics. After all, the intelligent document processing and classification set the stage for the next-level visibility and analysis of contract details that a contract analytics solution affords! Contract analytics is the use of instantly ready AI to analyze legal documents and their related counterparts to identify & extract important information from them in seconds, dramatically reducing contract review time. This leads to better risk management, faster due diligence, and tremendous efficiency gains – just to name a few benefits realized. Read more on the blog, “A Guide to Contract Analytics to Improve Business Outcomes.”

Right out of the box, an effective contract analytics solution should be able to use its document AI model(s) to parse the contents of a document and identify key points of concern and attention, including, but not limited to:

  •  Acceleration clauses
  • Arbitration clauses
  • Maximum liabilities
  • Irregularities
  • Non-standard language
  • Auto-renewals
  • and Evergreen clauses.

Moreover, the AI powering these document interpretations should be continuously improving, ideally without the customer organization needing to invest heavily in the training and maintenance process.

3. Intelligent, Contextual-Based Search

Super-fast search is a must-have in a modern document understanding solution if you want to ask deeper questions and get answers in real-time. The amount of data exposed by powerful AI solutions, while highly useful, can often present a challenge to using traditional search techniques. With a contextually aware search engine, documents that might have stayed buried indefinitely can instead be found with just a few mouse clicks, condensing work that might have taken weeks or months into just hours or minutes. The search functionality in a document understanding platform is not only fast but also context-based. This means it uses AI and ML models to correlate facts, obligations, and dependencies across documents. The additional context allows searchers to understand how documents relate to each other. This holistic and easily findable knowledge, which evolves alongside the documents themselves, supports faster and more informed decision-making. It also does not impose the burden of increased resources spent on manual review or building a custom AI solution.

4. Integrations and APIs

With a better understanding and visibility into document content, putting that data to work across your business for the various consumers that need it requires an ability to integrate with a wide range of 3rd party systems and applications. A document understanding solution may connect to other systems for:

  • Data warehousing.
  • Land asset management.
  • Enterprise resource planning.
  • Business intelligence.
  • Customer relationship management.
  • Estate and facilities management
  • Contract lifecycle management.

The most advanced Document Understanding platforms on the market will have pre-built Integrations with drag-and-drop configurations for maximum simplicity. For instance, a user just selects one system (such as your ERP, or CRM, or Data Warehouse) and connects it to the document AI platform in a few clicks— and/or available through a documented API. The end result is a versatile solution that can generate more ROI and business impact.

As a whole, document understanding offers the most reliable, economical way to classify and extract data from documents. For documents such as a lengthy memo, loan agreement, or contract, AI-powered document understanding saves immense time and money compared to searching for all of that unstructured data manually.

Document Understanding through automation and AI

By its nature, document understanding is an automated workflow. The embedded OCR and AI models powered by ML/NLP capabilities within the solution, systematically perform a set of tasks that would have otherwise required manual human intervention.

Let’s say an organization in the financial services industry needs to upload a collection of loan agreements and extract DSCR and DSRA requirements from them. Or perhaps a renewable energy company wants to understand how a change in regulation impacts all of its assets to determine if action is required or not.

For the end-user, the process is quite simple with a top-notch document understanding platform:

  • Easily Import Documents
  • Automatically Process & Extract Structured Information from the Unstructured Documents
  • Instantly Access Full Documents or Find Specific Information
  • Review, Validate, and Collaborate
  • Connect Document Data to Other Systems

[Learn more about how ThoughtTrace works!]

[Request a demonstration to see for yourself!] 

For readers interested in more of the technicalities, here is a bit more information about how a domain-specific document understanding platform works:

  1. The documents in question are easily uploaded via the solution’s intuitive user interface, which can accept a wide variety of file formats. For larger workloads, well into the tens of thousands, or for more automated integrations the API can be used to upload directly and securely.
  2. Document processing begins with OCR, which transforms the characters within the document into machine-readable text that can be indexed and searched by the platform. In addition to OCR, the form and format of the document is interpreted to infer how the structural elements of the document (bulleted lists, tables, etc.) inform the interpretation of it.  This information in combination with the text is used to improve the accuracy of many downstream processes.
  3. The AI models then go to work parsing the context and meaning of the data. Documents are automatically organized and categorized based on how these well-trained models interpret what’s in the text. For example, AI and ML might put a group of documents all pertaining to a specific subset of obligations into the same classification.
  4. Simultaneously, this real-time analysis of the documents yields precise metadata about them, including key dates and clauses for review. This data can be used immediately for search, discovery, and automation or can be passed on to a human team for review and for retrieval via intelligent document search as necessary.
  5. Subsequent intelligent document searches are both fast and contextual, with the ability to find information buried deep within documents and to highlight relationships between documents. If a contract’s implications change because of another document being uploaded and analyzed in the platform, the document understanding solution can help to
  6. Over time, processed documents can also be exchanged with and used by other systems, including those designed for data warehousing, managing land asset information, processing business intelligence, and more.

The underlying document AI and ML models enable all of these steps to be completed much more quickly than what’s possible with traditional workflows. Legal teams and other business users no longer need to resort to time-consuming processes built around manual review, complex file systems, or homegrown AI solutions.  Today, Document understanding solutions like ThoughtTrace come complete with all of the functionality & trained intelligence necessary to deliver value on day one.

Uses cases and advantages of intelligent document understanding

At a high level, intelligent document understanding expands the knowledge of what is contained within an organization’s documents, while saving time and being much more holistic than common alternative workflows. But what specific use cases do organizations look to address — and what advantages do they hope to gain — when using a document understanding solution?

Common use cases include:

  • Tracking renewals of numerous vendor contracts. Avoid overcharges and unclear or unfavorable contractual terms.
  • Reviewing documents to find underperforming assets.
  • Understanding leases to see how they’ll be affected by certain events.
  • Performing due diligence before a merger or acquisition.
  • Optimizing revenue and reducing liabilities.
  • Accelerating contract lifecycle management
  • Automatically populate a debt compliance calendar with your tasks to maintain compliance with lenders.

Modern document understanding solutions are built and optimized for particular use cases in sectors such as energy, legal affairs, financial services, healthcare, manufacturing, telecommunication, and real estate. At the same time, document understanding platforms will deliver substantial benefits to any organization needing systematic management and extraction of data from contracts and agreements.

With that in mind, let’s look at the major reasons for upgrading to a document understanding system that can deliver detailed analytics and actionable insights.

A single source of truth for reviewing and validating information

The challenge: Multiple siloed systems have long been the norm when trying to find renewal dates, evergreen clauses, force majeure clauses, and any other important metadata within documents. From accounting software and expenditure worksheets to cloud-based folders and emails, the complexity of such a workflow drains productivity while increasing the risk of errors and oversights.

The document understanding benefit: In contrast, document understanding provides a unified repository with maximum transparency into the processed and analyzed documents within it. Contract analytics are quickly and easily surfaceable in one place for lightning-quick time-to-value plus a deeper understanding of the business (because of course, at their core, documents are the lifeblood of any business).

Automated, secure, and auditable document classification

The challenge: File folders and sticky notes, whether physical or virtual, don’t scale to the challenges of document organization and categorization. Enormous sets of PDFs with nondescript filenames, hidden in hierarchical folder structures, can gradually become too overwhelming to systematically review. Meanwhile, both cybersecurity and auditability can suffer from such an ad hoc approach to building a document taxonomy and managing it.

The document understanding benefit: Document understanding harnesses the power of AI and ML models to automatically convert files into machine-readable form, so users can quickly search and uncover information later. Built-in document intelligence accurately extracts common clauses, provisions, and data points. Because AI and ML are continuously improving as more documents are entered into the system, this categorization and organization become better over time, fitting a company’s particular document conventions and requirements. The solution also enforces cybersecurity best practices and makes auditability easier thanks to clear-cut organization/categorization and accurate search.

Day-one AI readiness for powerful search and short time-to-value

The challenge: Developing and training an AI model is difficult work. The initial training is both time-consuming and challenging, as is the ongoing maintenance required to ensure that the model keeps delivering relevant insights. There’s also the issue of user-friendliness, as self-developed AIs are often not built with the end-user in mind, they can take a toll on productivity even if they’re technically sound.

The document understanding benefit: A proper document understanding solution packages up a properly trained and continuously improving AI that’s ready right out of the box. It can immediately begin using its embedded capabilities, from OCR to natural language processing, to dive into even the most complex unstructured documents to deliver critical insights. Because of this advanced AI processing, searches are both quick and contextual.

Accelerated due diligence and risk identification

The challenge: Manual contract review is risky. By the time a legal team completes its work, its findings might not even be relevant anymore. And that’s before considering the possibility of key contracts and details slipping through the cracks due to misclassification or difficulties with search. Critical activities such as due diligence get slowed down as a result of these issues.

The document understanding benefit: Document understanding automates much of the review process. It does the heavy lifting of extracting data from text and presenting it in an intuitive, user-friendly way to humans. Accordingly, due diligence can proceed more quickly and risks get identified early and often in the process. Users don’t have to choose between being right and being fast – they have the time to review ALL documents, quickly, with an AI-powered assistant.

Automated and integrated data systems

The challenge: Keeping track of sensitive, business-critical information across multiple systems and disjointed manual workflows can easily overwhelm an organization’s teams. As contracts evolve and more siloed tools are added and/or updated to support business growth, these processes only become more trying.

The document understanding benefit: A modern document understanding platform can connect with systems of all kinds, often through convenient pre-built integrations. APIs provide further opportunities to link the document understanding solution to other sources of information such as proprietary systems.

Conclusion

Document understanding delivers fast, intuitive insights. Unlike manual processes built around cumbersome file folders, a document understanding platform is automated and consistent, with AI and ML models that surface what matters the most and enable improved decision-making. Whether someone is looking for vendor overcharges or thinking about how to redo a lease, document understanding from ThoughtTrace offers a comprehensive solution that works on day one. Learn more by requesting a document understanding demo today.

Document Understanding Solution Infographic

 

.

Sign Up for News & Updates

thomson-reuters-acquires-thoughttrace