Tell us, what bot did you build with Josef?

We built a Q&A bot around unemployment insurance benefits in New York. We have an unemployed workers project that helps clients access unemployment insurance benefits, and also provides representation to them if they’d been denied benefits.

A lot of the questions that we get on our hotline are really basic: Am I eligible? How do I go about applying? Will my immigration status allow me to access these benefits? It’s a really high-volume hotline, and to the extent that we can get the basic information to clients without them having to call the hotline, it helps them get that information quickly. It also helps us in stemming the volume of calls that we have to make and frees up our attorneys and volunteers to do representation and other work as well.

How did you decide what to build?

It had a lot to do with the call volume. Clients were seeking basic information that could be conveyed through a fact sheet or a chat bot or a quick call. At the time that we were first approached about using Josef, this really was the project that had the biggest crisis with respect to the needs of the community and the volume of clients that we were seeing. 

We ended up creating something that was a step-by-step process for walking somebody through their eligibility for benefits. We use some visual tools, pictures of IDs and other documentation that somebody may need to show if their immigration status is a certain way in order for them to be eligible.

Before you automated this process, someone from your team had to have these conversations with clients in real time. How much time do you save with automation?

Each of those calls would range anywhere from 10 to 15 minutes. Very early on in the pandemic when people were getting laid off left and right, the New York State Department of Labor website was crashing because so many people were trying to access benefits. We were getting 20, 30 calls a day. We receive anywhere from five to 10 calls a day now with the automation. Even if a few calls are diverted using the bot, it saves our staff time, and we have a very small staff in this project, it’s just three people. To the extent that it saves them any time, it’s really awesome. 

Folks who are going to our website can click on the Q&A bot as opposed to calling us in the middle of the night when we’re not available.

At Josef, we’re really passionate about making justice accessible for all. That’s something you seem to share. Could you talk about the benefits you’ve seen for your community?

We’re an organisation that beyond serving clients, one of our main goals is to promote and foster volunteer work for attorneys. When Clifford Chance approached us about working with them, it was a great opportunity for us to connect with their associates and their summer associates that were working on projects through their innovation academy. This partnership did serve a purpose for us by promoting the work that we do as an organisation, promoting pro bono generally, and getting those private attorneys to understand the needs of the community.

And maybe they’ll come back one day and take a case or make a donation or do other work with us!

What was the actual process and experience of building the bot with Josef? 

We started [creating the bot] through this innovation academy that Clifford Chance held. Then we had folks from Josef come on to do some training. It was a really great and collaborative effort where our staff were involved too. I am not a tech person and I was still able to work with it and do some basic things and figure it out.

When [the bot] was handed over to us we got questions answered and had some support. It was definitely a challenge because we don’t have an IT person. We don’t have somebody who’s just building our website. We gave it to somebody that happens to do some data with us on our staff and said, ‘Hey, you want to do this too?’

So even though you don’t really have a dedicated IT team, you were able to figure out this process without that prior expertise?

Tech is always a challenge as a small nonprofit. Sometimes we don’t have the bandwidth and the resources internally to take on something new like this.

It was definitely challenging at moments, but it wasn’t coding. Like I said, I’m not the most tech-savvy person in the world and I was still able to figure it out.

Tell us: what’s next on your automation bucket list?

We’re going to try something with our immigration practice. Our case management system has a client-facing piece that you can put on your website, but it doesn’t walk you through it in the same way a chat bot would. So to the extent that we can marry those two things, I think that would be something we’d be super interested in.