Three of the biggest challenges associated with making any Knowledge Repository useful are 1) generating enough relevant content to keep up with the needs of its end users, 2) keeping that content up-to-date and fresh while underlying data and information continuously changes, and 3) enriching that content with advanced features that are well beyond the capabilities of manual content authoring, such as data visualizations. At the heart of the problem is the reality that manual authoring, formatting, curation and publishing of content is simply too slow, error prone, and expensive. As a result, many enterprises are turning to automated content generation solutions that address all three of these problems.
Understanding Automated Content Generation
Automated Content Generation (ACG) is the practice of using fit-for-purpose software that runs on computers to automatically create, format, curate, and/or publish knowledge articles.
If you’re familiar with traditional Content Management Solutions (CMSs), like Wikis and other web site builders, you understand that, in order to populate them, you must manually build, publish, and maintain your content one article at a time. This means there is tremendous reliance on humans, which is a bad thing.
Smart Knowledge Managers understand that work environments are changing so rapidly that attempting to manually document everything the enterprise knows is simply not viable. This is usually because 1) most employees have day jobs and don’t have the time or leadership direction to document what they know and 2) many employees believe that what they know is their own competitive advantage so they do their best not to share it with others, for as long as they can. For these reasons, far too many CMS-based Enterprise Knowledge Repositories (EKRs) have little useful content and, often, a great deal of it is out of date and wrong because it hasn’t been updated to keep up with underlying changes. And, for these reasons, smart Knowledge Managers are moving to tools that perform ACG.
The arguments for using ACGs are very clear and simple…
- Quantity: In the time it takes a human to manually write, format, curate and publish a single article, an ACG can generate hundreds of thousands or even millions of articles.
- Quality: The time it takes to organize and inter-link a knowledge article against and with many other related articles and web pages or the time it takes to go back and change all of it when underlying related content changes is very significant and only grows as more related content gets put into an EKR. This leads to all types of poor quality issues, like poorly classified content, poorly organized content, and many dead HTML links. ACGs improve and often eliminate most of these quality issues.
- Cost: The cost of funding many manual content authors is very expensive. ACGs are a small fraction of the expense.
Just how fast and thorough are ACGs? The Knowledge Management Body of Knowledge (KMBOK) is a knowledge repository that contains tens of thousands of web pages, hundreds of thousands of HTML links, and many thousands of embedded complex data visualizations. It was automatically generated in just a few seconds.
Automatic Content Generators Generate Far More Than Just Content
While the automatic generation of massive quantities of articles is clearly valuable, by itself, the real value of ACGs comes from their ability to go far beyond just article generation. The best ACGs also curate your content (e.g. classify and organize it) and they generate vast arrays of powerful Knowledge Constructs (KCs) (a.k.a. Knowledge Artifacts / KAs). Some ACGs even synthesize complete user experiences, where they weave all Knowledge Constructs together, into knowledge repository web sites called Digital Libraries.
Examples of synthesized Knowledge Constructs and Digital Libraries include but not limited to:
- Interactive and interlinked Library Catalogs, Indices and even entire Digital Libraries,
- Automatic content formatting (e.g. color coding, bold face text and templating of content),
- Sortable and interactive Lists, Arrays and Tables,
- Interactive Charts and Graphs,
- Interactive Reports,
- Interactive Dashboards,
- Different views of the same content (e.g. for different End User contexts),
- Interactive Data Visualizations,
- Vast quantities of bidirectional HTML links,
- Semantic Relationships and Semantic Data Graphs (SDGs)
- …and much more!
It’s difficult enough to get human content authors to take the time to simply document what they know, let alone getting them to enrich their content with complex formats, embed advanced Knowledge Constructs, classify and organize their articles, and then integrate them into complex structures like a Digital Library.
How Automated Content Generators Work
ACGs are, more often than not, Data Compilers. These Data Compilers work on a paradigm called Data Driven Synthesis (DDS), which ingests data and, along with some simple rules specification, work to automatically generate vast quantities of knowledge articles (i.e. content).
What makes ACG-generated articles notably different than human-generated articles is that they can also perform many advanced functions that are far beyond the capability of the average human content author. Also, ACG-generated content articles use a paradigm called Short Form Notation (SFN) versus the more traditional Long Form Notation (LFN) for content generation. LFN represents an essay-like narrative that uses long paragraphs to describe something. SFN uses short succinct Name/Value pairs to isolate and organize key attributes and their values.
Short Form Notation (SFN) is intended to get right to the point, eliminating the need to read through long, inconsistently formatted articles.
For those Knowledge Managers who will instantly argue that there is value to LFN… of course there is. However, LFN is expensive to generate and difficult to keep consistent, from article to article. Therefore, LFN should be the exception for article generation and not the rule. Also, LFN can easily be embedded within SFN attributes, ensuring the best of both worlds.
The Fuel for Automated Content Generation is Data
ACGs work on data. The data can from many different technical sources like spreadsheets, NoSQL databases, and proprietary Nth Normal Form relational databases. In an enterprise, think of this data coming directly from the specialty systems that people work in and around. For example:
- Application Management and Inventory Systems
- Capability Management and Inventory Systems
- Data Management Systems
- Facilities Management Systems
- Human Resource Management Systems
- Sales Management Systems
- Organization Management Systems
- Process Management Systems
- Product Management Systems
- Project Management Systems
- Technology Management & Inventory Systems
From a Knowledge Management perspective, these specialty systems become the collaborative Sources of Truth (SoTs) that context-specific Communities of Practice (CoPs) work in and around. For example, CoPs that care about Products work in and with the Product Management System, treating it as a SoT for most, if not all, Product data. Therefore, it becomes very simple to take extracts from these SoTs, such as an inventory of all Products with all their descriptive attributes and values, and push them into ACGs to generate Product specific articles, web pages, and knowledge constructs.
Automatic Generation of Knowledge Structures
As part of automated content generation, Automated Content Generators (ACGs) generate/synthesis massive quantities of Data Structures and Knowledge Structures that help format, organize, and convey content in many different forms, specifically with the intent to make it easy for humans to discover, learn from, and understand content.
Example Uses Of Automated Content Generation
The most prolific example of ACG use is in the development of an enterprise’s Intranet. Most Intranets have very limited content that is poorly and inconsistently formatted, classified, organized and inter-linked. And, more often than not, the is also very stale because humans simply don’t have the time to go back, find, and correct everything that is out of date. ACGs make it very easy to build content and feature rich Intranets that serve up-to-date content, all in a fraction of the time and in a fraction of the cost that it takes to build them with traditional CMS tools. (See an Example of an Intranet that was generated using Automated Content Generation.)
Another example of use ACGs is for the creation of context-specific knowledge repositories that serve a context-specific Community of Practice (CoP). For example, you might want a Customer Support Knowledge Repository that serves the needs of Customer Support Representatives or a Ship Building Knowledge Repository that serves the needs of Ship Builders. (See an Example of a Context-Specific Knowledge Repository that was generated using Automated Content Generation, specifically to support Knowledge Management Professionals.)
An Example of an Automatic Content Generator
The IF4IT NOUNZ Data, Information and Knowledge Management Platform is a Data Compiler that is used for Automated Content Generation. In addition to automatic generation of vast quantities of content articles, it also generates massive quantities of complex and interactive knowledge artifacts that, along with all articles, are all neatly classified, organized and interlinked within a massive Digital Library that is based on the architecture of a traditional brick and mortar library.
NOTE: Not all ACGs generate Digital Libraries or follow the same architecture patterns. The IF4IT always recommends that you evaluate as many ACGs as you can and appropriately match them to the needs of your own enterprise before selecting one.
Summary and Conclusions
ACGs are faster than humans, they yield higher quality, and they cost a fraction of a human workforce. Based on these and the above arguments in favor of ACGs, and based on their industry-tested performance results, if you’re a Knowledge Manager who argues against them, you might want to accept your fate and know that your competition is already embracing ACGs and, sometime soon, you will more than likely be replaced by them if you don’t embrace them, too.