June 12, 2018 | Written by: Jason McRoy
Categorized: AI/Watson | Social Service
Share this post:
I like to say that in my role as the Operations Division Director at Boulder County Housing and Human Services, I get to be a nerd who cares. I use my technology background and expertise in an organization that makes a real difference in people’s lives. So, I’m interested in exploring any technology that can help our staff provide better and more thorough assistance to the people we serve.
Dealing with information overload
Boulder County Housing has only 12 front-line customer service staff that see approximately 36,000 people a year. Our research shows that 60 percent of the clients we’re serving have two or more years of history with either our organization or another community-partner organization. That means that most people stepping through our doors have multiple areas where we can help. However, our staff needs to assess each person’s need and find a solution within a 30-minute timeframe. Often they only have time to address the most immediate problem, whether that is homelessness or applying for food stamps.
Our caseworkers and case managers spend a lot of time and put a lot of effort into entering very detailed, quality notes into our clients’ files. There’s a lot of rich information in those files about the person’s problems, what has been tried and what has failed and new issues that might have emerged. The frustrating part is that, once those notes are submitted to whichever database is relevant, it’s as if they become locked in a box and there’s no way to get any insights back out of them.
The only way that our client-facing staff can gain a deeper understanding about a client’s needs is to spend an hour or more reading all those detailed case notes. Unfortunately, given the volume of people coming through our doors, this isn’t possible to do for every client.
Our challenge was to figure out how to maximize the time that 12 staff had to serve 36,000 people annually in only 30-minute appointments. We wanted to provide more comprehensive services to our staff without them having to come up to speed and read the entirety of each person’s case history.
Unlocking critical information with Watson
As the technology guy, I was pretty excited when I saw how IBM Watson can understand unstructured text and present information to our customer service staff so that they can make better and more informed decisions for our clients. I thought it could be a great opportunity for Boulder County to add a layer of value to our client services.
Watson can read the case notes from seven different systems, identify risk factors and a person’s need for services, and provide that information in a scorecard to customer service staff. Staff will have that information before the client even comes into their office. They’re able to use this insight to advise the client and make recommendations and provide referrals to community agencies for services that they might have missed because they didn’t know what the real needs were.
Watson helps us see the entire narrative of each client at a glance. It can read and interpret a case file, which can be thousands of rows of comments long and that would take a human an hour or two just to read, in a matter of seconds. Watson essentially provides a 100 percent improvement in staff productivity without any additional manual effort on their part.
Helping families, impacting lives
Under our previous system, we missed opportunities to help people with their larger problems because we had to focus on immediate needs. For example, someone who comes in to find shelter because they’re homeless probably has other needs, such as mental health problems or joblessness. In the urgency to find them shelter, the customer service staff person may not have had time to delve deeply into the case file to identify and help resolve those underlying fundamental issues.
With Watson we’re much less likely to miss these opportunities to help our clients. We can now understand the whole person, and be able to make more targeted and accurate referrals to the resources that can help them. I think one of the biggest benefits, which was perhaps unthinkable previously, is that we can finally capitalize and benefit from the huge volume of case note data that we’ve captured in all our different systems.
Over the long term, we’ll be able to help stabilize and keep families thriving and functioning in the community. We’ll be able to resolve issues in a timely manner so they don’t spiral out of control to where the person needs more intensive and expensive services down the road for unmet or untreated needs. As the technology nerd in the organization, this achievement helps make my work more rewarding and worthwhile as I see the impact that using Watson can have on the lives of our clients.
Hear how Boulder County is using IBM Watson Natural Language Understanding to provide better and more comprehensive social services to its citizens by watching the video interview below with Jason McRoy: