Every once in a while we come across a person – or in this case, an organization – who is using these great technological advances that we marketers have access to – to advance something bigger than themselves. Today we’re talking about (and with) Code for America. If you’ve never heard of them it’s likely because you’ve never needed to. And that’s probably a good thing.
In short, Code for America has assembled a very nifty marketing technology stack – using Intercom as its core – to facilitate interactions between some of California’s most marginalized, least technically-adept people and that state’s service bureaucracies. As you’ll see, it’s no small task.
I hope you’ll also see the power that can be harnessed when great creativity meets great passion, meets great technology.
Thank you to you both for joining me. Code for America is on a big, difficult mission to use technology and human interaction to solve some of the heaviest bureaucratic issues facing underserved Americans. No small task. If you would, talk to us about your roles there.
I’m Elena Fortuna, I am Director of Client Success at Code for America. My name is Andrew Xie and I’m a Client Success Advocate with Code for America.
I guess something that’s a bit different for my role compared to Elena is that I do actually spend some time still conversating with clients, which means I still get to talk to folks, get a sense of what’s happening in the Calfresh ecosystem and also being able to provide resources and answer some of those questions in real time.
Explain as best you can what Code for America is – what you do. And then I’d like to talk about how it’s done and maybe more importantly at least from my standpoint right now, is why you’re doing it? It’s a really unique approach to some really unique problems.
So, Code for America – we have a lot of different projects that we work on. Our core values are to listen first, to really be driven by our clients and what they need, and so we have many stakeholders that we work with. We work with government partners and empathy is at the core of everything we do. So, we really want to try to include those that have been excluded, reach out to the most marginalized hardest to reach communities, and really make sure that they are getting dignified services and care. And so, a great way that we do that is by partnering with the government and supporting them, working shoulder to shoulder, to really make sure that the services that are being provided are easy to access and seamless in everything that we provide.
In terms of the actual chat service itself, I think based on like client experiences we’ve heard, research done by other parts of the team, what we have found is that a lot of folks that apply through us – through GetCalfresh – have had experiences with other parts of the bureaucracy. This means applying for other benefits or making use of other government services and for the most part, the experiences have been less than ideal.
And so, with the chat services that we provide with Code for America, in addition to just being supportive, simplifying the processes, giving folks confidence that we’ll be here, and we’ll ensure in the shortest way possible to get them the help they need – that we’re also just willing to listen. Some people just come in and in addition to requesting assistance, they just want somebody to hear them out, provide support, and just be there for them. And I think that’s also something that we do here that’s really unique, that’s probably not found in many other online chat services.
You kind of answered one of my questions which was – how much of this is automated? I mean, chat services have proliferated – mostly I’d say in chatbots – you can find them all over the place. But it seems like your approach is very human-centric, human-oriented, human-driven – you know, one person talking to another. Do I have that right or is this assisted by chatbots and automated chat services or how does it all work?
We definitely are. We are client-centered – our staff work directly with clients. So, we provide direct service, and we do have some features that Intercom supports us with like chat bots, auto closing, things of that sort. But we really want to be personalized in everything that we do and so everything – every response that we give – we really try to make sure that we’re addressing what the client’s saying. And especially because we have a lot of clients who are dealing with trauma – they have many different things that they are dealing with – complex issues. And so having someone there who can really walk them through every step of the application process, really provide that warm handoff if they need resources elsewhere, is really important to us at Code for America. And i come from a background of direct service and I know how difficult it is for clients to get through to somebody, and so we really try to be that listening ear and be supportive as much as possible.
I’m really intrigued with how your business model works because what you’re talking about is a very work-heavy model. You know, you can’t just automate empathy and have it really take care of what it needs to. So, how big is your client base? How do you manage all that? I mean, how is this done because it seems like there’s kind of a miracle of human resources in the background somehow.
Like you said, you can’t automate empathy. And that’s something we take into consideration whenever we try to implement any tools from Intercom and any outside sources. So, while we have utilized one of Intercom’s bots, which is called Resolution Bot. And how that works is when somebody first sends us a message, these are bots that we pre-fill with information meaning that we’ll train it to respond based on keywords in a message we get or other phrases. And what happens after that is if somebody gets one of those responses, the bot will follow up by asking if that answered your question, or you know, simply if you need to speak with someone else. And from there they can choose to either end the conversation but also have the option to reply at any time to reopen the conversation.
And what that does is it allows us to sort of filter out the more direct, easy to answer questions and give staff more time to focus on more complex, individualized situations. Especially during the pandemic it was really highlighted – huge demand and a smaller team. Much harder to keep up. But there were also a bunch of other tools in Intercom that helped us sort of manage and work through that volume.
A couple tools that come to mind: Intercom allows us to save templated responses into Intercom, meaning we don’t necessarily send them out as is, they’re what we like to refer to internally as like a starter kit for a response – meaning, this is where we generally get the policy side of the answer correct to this. But this answer is always going to have to be refined based on each person’s individual situation. But even having that tool has helped us save a lot of time.
A couple of other tools that we’ve also used as well. We’re also able to screen share with applicants and so if somebody sends us a message over chat, and for example they have a question about the application – like how do I fill out this part of the application or if there is some sort of communication issue in between, we can request for them to allow us to screen share and draw on their screen provide guidance help them click through the application.
So does that all live within Intercom – that function? Or is that a separate database?
It’s all in Intercom. We actually have our own database but our engineers on the team have worked to integrate it into Intercom, which also saves us time because we don’t have to constantly flip back and forth between the data to find the information getting context on the conversation.
For the uninitiated – I’m one of them – what kinds of problems are we trying to solve? What are some typical issues that someone using the service might be trying to get resolution on?
I want to say for the most part, based on themes and trends we’ve evaluated across all the programs where we’re currently using chat, the main one has to be “What’s the progress of my application?” “How do I move this forward?” “How do I get my benefits?” And then in between that, you’ll get a lot of the bureaucracy entanglement issues which include “How do I reach someone by phone?”, “My case worker isn’t responding – what do I do?”, “My caseworker is asking for a specific document I don’t have and I want a replacement”, and all the smaller things that folks oftentimes get caught up on trying to move things forward.
And then outside of our direct service would be requesting help with the application or Calfresh-related services. We also get a lot of requests on “how can I get help with rent assistance?” “How can I find emergency shelter, emergency food, anything that could help me in between this time while I wait to hear back from Calfresh?”
So in addition to the sheer quantity of people that you have to have manning these requests, they have to be fairly expert at navigating a complex system or several complex systems, which seems to me that, the more we talk about this, the more impressed I am by your business model and that you’re actually even able to pull this off. On no level does this seem like an easy thing.
It definitely is not easy. And just to clarify – we don’t actually have a robust client success team on the GetCalFresh team. We have about five people that are in Intercom throughout the day so you can imagine the volume that’s coming in. Thousands and thousands of messages on different channels, whether that be live chat, SMS, emails. And so, a big piece of our model, and what we are moving towards, is actually bringing in the caseworkers and training them on the best practices, and how to use Intercom, how to use email and these other pathways of communication other than just phone calls. So that’s a model that we’re moving more towards because we aren’t a client success agency – that’s not our model. But we need these proof points, we build out these applications – these platforms, and then we man them – we staff them in order to get these proof points and to support our clients.
You just answered my next big question – is this a transferable model with the technology intact or is it kind of a start-from-scratch effort when you show up in a new state? And I guess from what I’m understanding, it’s fully dependent on the systems in that state but the chat system seems like it’s a model that you could basically pick up and replicate elsewhere.
That’s something that we’re working on is building out playbooks and toolkits to really support any state to bring this on – to really build out their own client success team and be successful. Because a big piece of what we’re focusing on too, is staff support. You can imagine the high volume for our staff – also there’s a lot of vicarious trauma they’re dealing with – situations that, you don’t know the end game of what happened. So, this client might come to you and say that they’re in a really difficult situation with somebody, maybe there’s domestic violence, maybe there’s something else happening, we do a warm handoff and then we don’t know what happens next. So there’s a lot of support that staff need as well which, I think a lot of government agencies don’t always prioritize. That’s something that we really want to partner more with them in doing, is supporting the client and the staff and being holistic in that perspective.
You’re probably aware of this but there’s a government mandate, basically across all agencies, to improve customer experience and we’ve published a little bit on that. But you guys seem to be right in the sweet spot in a way, to help agencies of all kinds manage customer experience issues.
So, thank you to both of you – Elena and Andrew – for your time and for your input. Best of luck with Code for America. It’s an awesome idea and it seems like you are actually executing something that for mere mortals, would be next to impossible. So again, thanks for your time. We’ll be paying a lot of attention to what you’re doing.
Photo by Jonathan Bell on Unsplash.
2 comments