Maia is a new virtual assistant that Khoros built to help our customers directly from within Atlas, and throughout Khoros' website(s). It's the culmination of years of research and development around how to best help users adopt and utilize Khoros products.
Maia's name is rooted in Greek mythology, where Maia is the daughter of Atlas and the mother of Hermes (the messenger of the gods). Maia was designed from the ground up to reflect Khoros' company values, with a voice and tone crafted by Natalie Houchins, a talented member of our Information Experience team.
Maia is powered by Khoros Modern Chat's Automation Framework, and acts to both direct users to the answers they need or to one of our Atlas Guides, our dedicated chat agents. Maia is capable of filing support tickets, checking product status, and much more.
So, what went into the Maia project? What type of technologies are powering it, and how can you leverage these solutions to create your own interactive chat experience for your customers?
A Discussion with Maia's Creators
In order to tell the story of Maia, we talked to Travis Berryhill, the leader of the Maia project, and Anshul Jain, the engineer that led in Maia's development.
What inspired the creation of Maia?
Travis Berryhill: For the longest time, chatbots just weren’t up to task when it came to actually being helpful. So, we wanted to deliver a chatbot that would actually be helpful. A bot that could complete tasks for users and peers and truly make their lives easier.
What are some of the technologies the team utilized in Maia's creation?
Travis Berryhill: I like to use Lucidchart to plan out the flows before we get started. We use Flow.ai to actually build the bot and train its Natural Language Processing (NLP). We have utilized several APIs, like Statuspage and Salesforce in order to help Maia accomplish tasks.
Maia is powered by an amazing startup called Flow.ai, which happens to have just joined the Khoros family mainly because of the superior bots they were able to produce. Flow’s artificial intelligence and machine learning functionality is incredible, and it gives Maia the potential to be a game changer when it comes to bots.
What made Flow.ai the platform of choice for Maia?
Travis Berryhill: There are so many reasons! Flow.ai had the cleanest UI out of our choices, so that was what caught me at first glance. Then, once I dug into the functionality, I was blown away.
Its artificial intelligence and machine learning capabilities are incredible, and it had Maia quickly discovering intents that weren’t even programmed. Flow is also continually innovating. It feels like it has new functionality all the time, and it all just made our lives easier.
From a development standpoint, are there any challenges that you faced when creating Maia? How did you overcome them?
Anshul Jain: Initially, there were a couple of challenges to understanding the process and how chatbots work, but Flow.ai's great documentation and the Khoros Community Developer Docs made life easier.
From a technical point of view, whenever I find any issue either with the code or in the flows, I just check the documentation. Most of the time, I find my solution. Occasionally, I get issues while calling the APIs, so before using any API, I test that in Postman first and then use it in chatbot.
What advice would you give a developer tasked with creating their own interactive chatbot for the Automation Framework?
Anshul Jain: I would suggest starting with simple flows. Once you get a mindset on how the chatbot can solve the manual process, then you will start getting more ideas.
To develop the chatbot, I think that the developer should be proficient at JavaScript and problem-solving. And before starting any development, write down all the steps. Once you are clear with the steps, then start the development.
Is there anything exciting coming to Maia we can talk about?
We are about to release a flow that will allow users to open support cases from wherever they are instead of having to go to Atlas, every time. I’m really excited about that.
Development Nitty Gritty
Maia is powered by Khoros Care and its incredibly versatile Automation Framework. The Automation Framework enables you to connect chatbots to Modern Chat and facilitate direct interactions with customers, as well as to hand off those interactions to live agents when the conversation calls for a human touch.
Adding the words and interaction trees to Maia was accomplished using Flow.ai. Flow handles the message responses throughout the entire automated conversation. It also initiates the handoff when and if it's time for a human to step in.
The actual connection between the chatbot and the customer is all handled through Khoros Care and its Modern Chat. The Automation Framework API facilitates the exchange between Flow.ai and the customer, and then hands the conversation off to an agent when Flow.ai triggers the handoff.
From there, an Agent signed in to Khoros Care can interact directly with the customer in one seamless experience.
Maia's Incredible Team
No discussion about Maia is complete without recognizing the many folks that helped make Maia possible. In addition to Travis Berryhill (TravisB) and Anshul Jain (AnshulJ), a whole host of people contributed to Maia's continued success. These folks include:
