Let’s start with you and government leaders like you. What are we hearing from them? Around the country, it's all about how do we do more with less money while continuing to serve our constituents. Even more so now, we continue to serve more constituents with flat or decreasing budget dollars. We’re constantly looking for ways in which we can serve more people with the same amount of money and provide exceptional services.
Our goal with the video above is to share ideas about how to drive down costs and how to help you provide solutions that benefit your citizens and keep your staff from being overworked.
Google Cloud is a platform consisting of software as a service technologies (SaaS), platform as a service technologies (PaaS) and infrastructure as a service technologies (IaaS). You may be familiar with some of these concepts, but here we focus on a platform as a service technology called Dialogflow. What makes it unique?
Speed to value & burden reduction
With PaaS technologies, you have the ability to achieve your goal using an application and roll it out without having to worry about infrastructure, compute storage or the plumbing that typically goes along with creating an application.
Dialogflow helps government organizations as well as commercial organizations interface with constituents.
Let’s look at an example involving 9-1-1 calls. In cities like San Jose, technology like Virtual Intelligent Agents (with Dialogflow) take about 20% of calls from human call takers so that they can focus their time on more delicate and complex types of calls. The technology ensures that if it's a burning building it's going to 9-1-1. If it's a burning question, it goes to 3-1-1.
State and municipal governments use Virtual Intelligent Agents to help have rich conversational experiences with their constituents and to automate some of the workflows that are burdensome to the call center.
These virtual intelligent agents are designed to be assistive technology to your human staff. They should be able to solve common problems through websites, mobile applications, popular messaging apps, and be embedded and communicate over the phone system on your web page, through twitter, facebook messenger, and a variety of other platforms.
High vs. low-value tasks
Dialogflow is powered by Google's Machine Learning, which is important because machine learning can be assistive in helping you with routine conversations that you might have with your customers. If you think about the support that you give your customers, there's a variety of low and high value tasks, but in most cases it's the same people that are delivering both types of tasks.
What if you could take your high value people and enable them to focus on the high value tasks while providing a service to automate some of the lower value tasks?
Some examples of these everyday low-value tasks are things like password resets, claim statuses, account balances, someone forgetting how to log in, questions around tax deductions or how to file for a picnic permit. A lot of questions come in and the answers probably already exist on your website. The problem is, in most cases, people have to hunt for them. If the answers are not readily available, you can have a virtual intelligent agent there to say hello and help navigate them accordingly without getting additional telephone calls from frustrated users.
The need for simplicity in a self-service world
Studies have shown that up to 80% of customer interactions can be resolved by well-designed chatbots. Consumers today want to self-serve. Very few people really want to spend the time on the telephone talking to a human. As people become accustomed to self-service for things like retail that are giving them such an excellent user experience, the government can do the same thing.
In government, where you have periodic surges (like during tax time), many departments of revenue bring in a number of seasonal workers to man the phone lines. Especially today, across the country, unemployment and departments of labor are struggling, citizens need help, and therefore the government is staffing up with temporary workers to handle the surge.
Technologies are becoming ubiquitous in terms of helping people. Studies have shown that in the future there will be a lot more investments in these technologies because of the value they bring.
What is seen across the country with state and local governments is that the universe of applicability falls into three major categories; connecting agencies to their constituents, increasing accessibility and automating routine tasks. Let’s take a look at each.
Connecting agencies to their constituents
Constituents are both internal employees and external customers. The help desk is a really common use case for this technology, especially these days when people have been forced to work at home. There has been a huge surge of people not ready to work from home, facing problems such as lack of internet access, not being able to connect to the VPN, and trouble accessing applications and file shares.
Questions around these problems are also examples of low-value tasks. In this case, a virtual intelligent agent can walk your user through a diagnostic procedure and leverage the knowledge base that you already have.
From a customer service standpoint, let’s revisit the unemployment example. These questions may include: how do I file a claim, what's the status of my claim, what are your office hours, how can I change my payment method, and the list goes on. Virtual agents are quickly resolving issues for these external customers while also managing your internal employee help desk..
The second area we see virtual agents being applicable for is increasing accessibility. Virtual agents, using Google’s powerful Machine Learning translation, are multilingual, speaking over 20 different languages. This allows you to provide support to your constituents in the native language they speak. With many local ordinances and state regulations requiring information be made available in up to six different languages, virtual agents with Google Dialogflow are at the rescue.
Automating routine tasks
Automation was brought up earlier when we introduced common low-value tasks, such as pin, account and password resets, status updates, etc. With most of these answers in your systems already, Dialogflow agents can connect to your back end systems and automate the delivery of this information.
Knowledge delivery is also important when bringing seasonal workers, new hires and/or temporary workers up to speed. You can take the knowledge that you have in your systems and train a virtual agent to be able to deliver that knowledge through an assistive technology to your call center agent. They can ask IT questions and get answers to deliver to the human on the phone or over an email immediately.
But how do you train these virtual agents in the first place? Everything that the virtual agents converse with is recorded in a data warehouse. Once there, you can start running analytics on these conversations; what are people asking specifically, what are the most popular topics people are asking about, what questions are they asking that we can’t yet answer?... This last insight will allow us to enhance the virtual agent accordingly.
Virtual agents are the Tesla of chatbots
“Chatbot” is the buzzword that most people use to describe the idea of this technology, but by industry consensus, they don’t have a great track record. There are a number of reasons why they’re not very good. 1) They're not built with artificial intelligence, 2) they use keyword searching, and 3) you are forced to set up complicated if/then/else structures. You have to program almost everything that anticipates the way a person might ask a question or the topics they might ask, otherwise these chatbots fail.
This is quite different from a purpose-built intelligent virtual agent. Much like an old Ford car is not a Tesla, not all technology is equal either. Not all chatbots are equal. In our video above, we cover live examples across a variety of websites so you can see the differences of these technologies in action.
How are virtual agents being used during the pandemic?
Governor Pritzker of Illinois was holding press conferences everyday during the pandemic, after his press conferences his office would get raided with questions. To help, they posted a virtual agent on Governor Pritzker’s website in English and Spanish.
In the first two weeks it handled 3.2 million calls. Two weeks later we added the virtual agent to their phone system on the after hours phone line. Before that, there was no way to talk or get help after hours. Now about 40,000 calls are handled after hours with virtual conversations and the results are showing that there's a 94% success rate of the interactions with the virtual agent, which means people are getting answers that satisfy their questions.
Getting started is as easy as a phone call. Cloudbakers can help you build a virtual agent plan that gives you suggested use cases, implementation timeframes and ROI. Reach out to us today.Originally published on September 17, 2020