Organizing training documents and reference materials into a usable format is fun!
If you disagree, then you are normal. Almost no one enjoys gathering materials in different formats and different styles and then standardize it to make it easy for people to consume. The typical technical solution is to distill everything down to basic text files, store those in a database and call it done.
This does solve the problem of finding the documents and provides a means to search within them. The search process is more of a simplistic find function and doesn’t have any intelligence behind it.
This level of simple search doesn’t help employees or other stakeholders extract useful information. It is akin to simply putting everything in a single document and telling people to use the find function. The person can find all the locations of a target word, yet it doesn’t help the searcher understand the context of each appearance. The context and meaning of the section require more reading and time.
That has been the problem of knowledge management systems for quite some time.
Still, there is no denying that creating completely new material from scratch geared for a smart, interactively searchable system is not feasible. Organizations need to find a middle ground between which takes existing materials and ingests them into a system that then allows smart, context-sensitive searches.
Taking that concept to the real point of use becomes a micro-reinforcement system.
Micro-reinforcement starts with reinforcing training. Real-world use is more akin to an intelligent question and answer system. Essentially this provides an “easy to consume” method for building know how and improving retention.
We all know that people do not retain all the information provided in a training class. Some research finds most people retain less than 50% of the material after only 5 days. Other research isn’t even that optimistic. To pretend employees will know what to do or how to do it after just a training course is naïve.
Micro-reinforcement
Reinforcing what was previously learned is necessary. Nobody wants to admit the end of a training class wasn’t the last word in knowledge delivery. Frankly, most employees are ready to return to their regular tasks or other activities and meetings. Possibly, they are newly hired team members and are overwhelmed and need time to digest what they’ve heard or learned.
Answering specific questions after the training session is more palatable and necessary. This is micro-reinforcement. It is much easier for everyone to get behind as a tool in the organization.
Micro-reinforcement of previously learned information with immediately useful targeted information makes sense to everyone.
Start with Existing Materials
Micro-reinforcement is an evolving technology. Its intent is clear, and the purpose easily understood throughout organizations. So, where do you start?
Let’s start with the existing training handouts, process documentation, manuals, and other reference materials. If you try and convert them to a standard format and apply a search function, the result is barely more effective than having the materials scattered as they are now.
To be more effective, the existing materials need some processing – both advanced machine learning and artificial intelligence (AI) processes and a little manual tweaking.
Cracking
There are machine learning processes that can aid the ingestion process. Existing materials are almost always in one of a couple of handfuls of file types. Typically, most training materials are in PDF, Microsoft Word, Microsoft PowerPoint, text files, and / or other equally common file formats. There are standard methods of ingesting those documents.
This first step of bringing the documents into a micro-reinforcement system is often called cracking. Think of the material inside your documents. It sits in there nicely. To get the information out, you must “crack open” the document.
Enrichment
Enrichment is the concept of taking the cracked material from the various sources and making it truly usable for micro-reinforcement.
Small segments of material must be extracted. Keywords or questions need to be tagged for each section. Synonyms and similar words have to be associated and added to the base material to make it searchable in ways real people want to interact with the system.
Working with the cracked content to create easily searchable answers to specific and general questions in micro-reinforcement is enrichment.
Artificial Intelligence Ingestion
Cracking and enrichment go hand in hand. These are two steps that must be done to prepare the foundation of any micro-reinforcement system.
Fortunately, much of both sub-steps are made much easier by machine learning and artificial intelligence. There are numerous methods in machine learning to crack documents. This dramatically shortens the time needed to prepare and ingest. There still needs to be human curation of the cracked documents since not everything in all the source materials is useful.
Natural Language Processing (NLP) functions from AI methodologies aid in enriching the materials too. Being able to apply standard models of human speech to your source material speeds up the enrichment process 100 times doing it manually. NLP procedures produce a first pass over the materials that humans can review and edit. The edits are useful to further target and apply the vocabulary inside your organization.
Vocabulary
One of the most interesting aspects of applying NLP to cracked materials is how people have different vocabularies in different parts of their lives. People do not talk the same at work as they do at home or around friends. Different companies have their own vocabulary too. Tailoring and refining the enriched material to match how people in the company would understand and search is needed.
Generic NLP processes can get source material enrichment close and eliminate most of the work in extracting questions and relevant descriptive words. NLP processes can even create questions that employees may ask of the material.
This is a useful starting point, but it still needs human curation. Curation focused on altering the generic extracts or eliminating words and phrases that would not be used inside the company.
Deployment
At this point, the material is ready for deployment inside a micro-reinforcement system. Seeing how words are used in questions directs people and NLP processes to improve the questions and vocabulary to make it more accurate over time.
Without a framework and smart ingestion systems, organizations will never get through the step of ingesting materials. Micro-reinforcement needs smart extractions from original materials to be useful and impactful.
Payback
The smart micro-reinforcement system helps organizations in the short and long term.
Short term, employees do their jobs better and need less interaction from supervisors. This lets managers and other employees focus on their own tasks. There are fewer mistakes, less waste, and more productivity.
Longer term, the same gains hold, and the organization can continuously improve their training programs to match real world issues. Rather than guessing at where gaps in training exist, or tasks that are not well documented, the organization has a record of when and who needed assistance.
Investing now in ingesting materials and enriching them using AI for targeted consumption gets the smart organization into a powerful position for future efficiency and growth.