Custom Training Upskills Staff to Prototype Chatbot In 5 Days

Improving Access to Healthcare for 125 Million People

machine learning
Optum Health
machine learning
2018
machine learning
Healthcare
Optum Health logo over a doctor on a tablet speaking with a patient

At a Glance

With the goal of providing better access to healthcare information for over 125 million patients, engineers at OptumHealth were at a standstill: Their data-savvy team lacked the confidence to deploy robust automated solutions given varying skill levels team-wide. Data Society developed a custom training program to establish a baseline understanding of how chatbots could be implemented across OptumHealth. Empowered with new skills, the engineering team was able to prototype the “drug chatbot,” where end-users gain invaluable information in order to better manage their medications and overall health – all in five days.

Client Profile

Bio:
OptumHealth provides health technology services. The company offers data and analytics, healthcare delivery, information technology, revenue cycle and population health management, and financial services. OptumHealth operates worldwide.
Industry:
Healthcare
Headquarters:
minneapolis, mn
Revenue:
$100B
Founded:
2011

The Challenge

OptumHealth, a subsidiary of UnitedHealth, strives to shape a better healthcare system by bringing every aspect of the healthcare system together, giving 125 million people access to personalized health care services, including pharmacy and healthcare financial services. 

When OptumHealth approached Data Society, senior engineers at the company were developing chatbots to give its customers to have easier access to key health data. However, engineers found it difficult to teach foundational concepts of implementing chatbots across teams of varying skill sets.

The Solution

Data Society developed an on-site bootcamp for experienced programmers at Optum Health to establish a baseline understanding of how to use chatbots to achieve customer and business outcomes with Python and Javascript. The training covered the complete process of implementing chatbots from the development stage, different tools and platforms used in the building stage (including how AI tools such as neural networks identity patterns in text and conversation and use them to mimic human conversation to beat the Turing test), and how to program Facebook messenger chatbots.

An additional bootcamp was added specifically for software developers looking to implement their knowledge as soon as possible. The “Using AI and Natural Language Understanding (NLU) to Create Custom Chatbots” course focused on deploying chatbots in production environments and tailoring chatbot AI to a particular domain.

The Results

Like many of Data Society’s signature training programs, this course concluded with an applied capstone project. In this case, programmers worked with real-time organizational data sets to create and implement several different chatbots that can be used to solve specific use cases. Developers were able to conceptualize a “drug chatbot” that would allow customers to learn the active ingredients in drugs and which drugs may potentially have a negative interaction with other medications a patient may be using, helping solve for unintentional overdoses or negative reactions. 

As a result, the Optum team didn’t just imagine how they could improve the health and safety of their patients, they acquired the skills to continue making a positive impact across the healthcare system for years to come.
cross linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram