There are times when we are reminded why we work in technology and how the technology we develop and sell can make a meaningful impact. Today, I’d like to share a story that could have had a better outcome if Natural Language Processing (NLP), integrated into a Knowledge Management solution, had been applied. We at Pryon are helping field support organizations with that exact solution.
Let me start with a little background. I am fortunate to live in Charleston, SC, a beautiful city with its Live Oaks, great restaurants and charming provincial homes. Tourists flock to Charleston to witness southern architecture first-hand. However, what they don’t typically realize is that just below the surface of these centuries-old houses is infrastructure requiring layers of renovation and repair.
My neighbor’s home was built over a hundred years ago and, while beautiful, it almost always yields unwelcome surprises when a project is underway. This was true a few weeks ago when a tree was cut down on a neighboring property and took out the cable line that connected her house to the rest of the world. The cable company responded promptly to make the repair, but after 3 attempts to restore her cable service, the service techs left frustrated. The challenge: it would take an exceptionally experienced technician to understand the issue and how to solve it.
This is where NLP can make a difference – Helping field technicians understand challenges, accurately assess the issue and identify a solution. The technician could have (1) asked an NLP system a question about what they were seeing and (2) been provided with answers from a catalog of ingested content. How much time and money could have been saved, plus what would have the customer satisfaction score been if they would have solved it on the first visit? Further, how many times a week or month is this happening?
Frankly, I love this use case. It’s impact touches the technician, the company or companies, if there are contractors, and the customer. Applying NLP in the field can make jobs more efficient and profitable while keeping technicians safer. All of these benefits are typically core to a company’s goals and easily measurable.
The “how” is the question and let’s start with what is NLP. CIO magazine’s definition is the branch of Artificial Intelligence (AI) that deals with training a computer to understand, process, and generate language. There are sub categories that include Natural Language Generation (NLG) which focuses on the ability for a computer to create communication and Natural Language Understanding (NLU) which is the computer’s ability to understand acronyms, synonyms, slangs, abbreviations or other alternatives to a word. This is key to how NLP differs from Keyword Search (Keyword).
Let’s look at an example of Keyword versus NLP. In one of our engagements we are applying our NLP platform to help field techs support industrial equipment. Part of the engagement is focused on the safety of the tech while in the field. We are ingesting several hundred documents, some of which are nearly 1,000 pages long. Today these documents are in a shared folder that someone can search. The word “fire” shows up in over 50% of them and in one document it showed up 279 times.
Imagine being in a situation where you need to know what to do and searching through over 500 documents to find the answer to your questions. Now imagine doing this while stressed, maybe using Text To Speech (TTS) and your second language. Let’s try something, take a moment and write down all the ways you can ask “How do I get off the rig during a fire?”
I suspect you came up with a lot and Keyword is only going to get you to the document. Now you need to get to the place to answer the question. NLP measurably lowers the difficulty and time of finding this answer. To expand the question and answers understood by the system, we have added a dictionary. This enables a company to include company or industry specific terms or acronyms, synonyms and commonly mistyped words. Adding NLP to a system can save lives by making it easier for someone to find the answer to their questions.
Let’s go back to our example of the cable being out and everyone would like it fixed on the first trip to the house. How could NLP help solve this challenge? As mentioned above, the issue was solved by an experienced technician that had seen this type of challenge before. But, this issue was documented in the repair content provided to the technicians. However, a tech needed a few things to access the right file.
- Confidence that they would find the answer if they searched
- Knowledge that the answer provided was correct
- A system that provided easy access to remote workers
The difficulty was while they had a repository of documents they were not easily searched and the answer identified. It was far easier to either assume they had solved the problem or call another tech to get their assistance. This is what created the circle of multiple visits before it was ultimately solved by an expert. I can imagine (I did not see it) that the company has a checklist of tests a tech should run before they leave. Not just turn on the TV to see if you get a picture. In this case the experienced tech went to one of the junctions (remember this is a bit of a hodge-podge) and wiggled the wires there to see if the picture remained on. When it didn’t he knew that was the problem area and fixed it. The tree had made the connectors loose and with wind, a squirrel etc they would separate enough to lose the connection.
If the initial tech had asked “What else should I check before I leave?” they would have found the loose connection and been able to solve it. They didn’t not because they didn’t want to solve the problem or weren’t smart enough, they hadn’t seen it before and I suspect some of the kludged solutions in Charleston were not part of the training. But if they had been able to ask the question and receive an expert verified answer from their phone the problem could have been solved in the first visit. To the benefit of everyone.
Pryon’s NLP Platform is helping enterprises enable their field technicians to be more effective, efficient and safe while performing work that is crucial to their customers and their long-term profitability. We would appreciate the opportunity to talk with you about how we can work with your team to create similar results.
– Scott Hoffman