The Silicon Review
Three professionals who worked together at a medical services organization, Robert Marsh, Matt Berseth and Ted Willich, decided to start a new company focused on delivering natural language processing (NLP) solutions to the healthcare industry. Little did they know, they would soon pivot after winning a Kaggle data science competition, to delivering machine learning solutions across many industries and become one of the fastest growing AI companies in the United States.
In 2012, Lead Scientist Matt Berseth was placed 2nd of 146 in the Kaggle Practice Fusion classification contest. The contest identified patients with undiagnosed Type 2 Diabetes and launched the new business model of focusing on delivering machine learning solutions to industry.
Benefits of Automation in an Uncertain Market
With discussion of an economic downturn on the rise, the days ahead may seem as unsteady as walking a tightrope with the consequences leaving many performers asking if the safety net is secure and what other checks are in place? Pivoting to find balance, organizations are turning to the benefits of Robotic Process Automation.
Many organizations have already found early success in automating redundant but necessary tasks such as data entry, document processing, medical coding, as well as other time-consuming areas of their operations. These types of projects are often the first steps in an organization’s RPA journey as these types of successful projects open up other automation opportunities and expansions in current project scope. CFO’s are now looking at how they can use AI and Machine Learning to automate areas of their own department such as the month-end close process, collections, payables, and budgeting for additional time and cost savings.
The reason for the interest in these projects is simple – a one-time project cost that will return perpetual savings. Most of the heavy lifting for AI/ML projects is done up front. This includes gathering the data, building the actual machine learning model, testing the accuracy and putting it into production. Once in production, stakeholders are able to realize cost savings almost immediately and in many cases they are able to recoup their initial project costs within the first 3-6 months.
Q. Where to Start?
One of the first questions that are often asked when exploring AI and automation projects is where to start? As an organization, you have to identify the best use cases for AI so you can start off with some easy wins. Some questions you might want to ask are:
Sometimes finding answers to these questions can be challenging, especially in regards to the data. If you are struggling to find a good use case, have questions on your data, or just not sure what AI tools you will need for your project, NLP recommends starting out with an assessment before diving into your first project. An initial automation assessment can identify the best use cases for applying AI and give you an “AI project roadmap” along with cost estimates, project timelines and potential ROI for these projects. This will allow you to prioritize your projects and include them in your overall business plan.
Theme and Sentiment Analysis
Many organizations are focused more and more on the experience of customers, employees, and partners recognizing that behind every person is a story. These stories can help organizations illuminate and improve moments that matter most to individuals. Often these stories are gathered through open-ended questions in surveys and questionnaires.
As more and more qualitative feedback/stories are collected, the problem lies within how to leverage the insights, positive or negative, for continuous improvement in organizations. These stories can serve to assist organizations in recognizing the behavior, preferences, wants, and needs of those they serve—not as point-in-time insights, but as an ongoing relationship. “Businesses know there is richness in the qualitative feedback they have collected, yet many struggle with the best way to utilize it in a systematic way to help make better business decisions, quicker,” states NLP Logix Modeling and Analytics Lead, Mary Sheridan.
To bring an understanding to this mass of open-ended feedback NLP Logix has developed a theme and sentiment analysis solutions to gather, collect, categorize and inform organizations of insights. Theme and sentiment analysis was built utilizing Natural Language Processing (NLP) to do the following:
Sheridan states, “While many companies rely on more rules-based logic to classify feedback, NLP Logix uses several layers of NLP technology to dissect comments into meaningful units, extract relationships between words and phrases, and identify similar concepts that should belong to the same theme. This cutting-edge technology allows the model to grow and adapt to new concepts and without the need to manage large and complex rule sets.”
Utilizing comment themes and sentiment to identify opportunities to improve experiences, organizations can additionally track improvement progress after interventions have been put in place. For example, a key initiative for many industries is to ensure that organization leaders are providing all the information and education needed for customers to continue the proper course for success. Utilizing NLP Logix’s custom solutions, organizations can track the percent of negative comments about the information and education they are receiving, and work to turn those into positive comments over time.
“The NLP Logix solution allow businesses to uncover growing trends in real-time before they become a larger issue,” states Sheridan. “Hearing multiple customers or employees discuss a similar topic in a negative light can help draw attention to issue before it has wider impact.” As organizations have become more astute at utilizing feedback/stories, expectations have evolved. NLP Logix enriched the current NLP comment analysis to enhance the following features:
To solve these challenges NLP Logix has created a specific model update driven by NLP techniques enabling:
Priority Support for Ticketing Process
Regarding P1s, it seems suspiciously circumstantial that keyboards have the “1” and “!” on the same key. P1s are “Priority 1,” top-priority Support tickets addressing business-critical needs. If a support ticket comes in assigned as a P1, urgency from the NLP Logix Client Operations Team is actionable with the emphasis of taking action with an exclamation point! As tickets are channeled into the pipeline, they are assigned a priority level of P1, the most urgent, to P4, the lowest urgency. These levels are based on the ITIL standard if identifying the impact and urgency of the issue. Kristine LaBarbera, Director of Client Enablement, states, “We never look forward to a P1 scenario, but they are a necessary and anticipated part of operations. They also provide our teams with the opportunity to identify areas of optimization to prevent the same scenario from occurring in the future. We’ve been working to refine our process over the course of the last year, with some great input from our clients along the way.”
Eli Butler, NLP Logix Client Operations Specialist states, “When monitoring Client’s processes and health, P1s are not common. However, when a P1 does occur the entire team’s focus goes to handling that issue and we work seamlessly until that issue is resolved.” NLP Logix believes Data Science is a team sport® and carries that into Client relationships. They encourage Clients to lean on NLP Logix as an extension of the Client’s business. Throughout the P1 process, Clients are heavily involved through the remediation process. The P1 process is highly collaborative with various teams from NLP Logix and the client participating to drive resolution as quickly as possible. In alignment with the ITIL v4 Guiding Principle of “Continual Improvement,” NLP Logix strives to assess and improve operations, performance, and value. The P1 process is one piece of the full service of end-to-end delivery of quality solutions for our clients.
Ted Willich, CEO