In the past few years, bots have exploded in popularity far beyond Siri. Facebook Messenger now has 11,000 bots for users to speak with, from an Invisible Girlfriend that acts as a custom digital girlfriend to a BFF Trump bot that replies to queries with quotes from presidential candidate Donald Trump. Of course, bot software also automates more serious tasks, like recruiting workers for open job positions and monitoring customer interactions.
The advancing technology leads some, like Facebook founder Mark Zuckerberg, to assert that bots will largely replace human help within the next decade. However, we haven’t arrived there yet. Text-based bots still fail fairly often, which can lead to a negative customer experience, so bots still need to be supplemented with backup from live people.
If a bot fails 5 percent of the time, that means every 20 inputs, you’re getting one input that fails. Imagine if you’re speaking to someone and every 20 sentences, they respond back with something completely random. You wouldn’t be friends with that person very long. Likewise, customers probably wouldn’t stay with a company that did that either. Therefore, the best bots today combine software and human components for a superior customer experience, such as our bot, Overseer.
First developed about two and half years ago, Overseer helps companies using Glia engage millions of visitors to their sites more effectively. Based on where potential customers click, navigate, and otherwise move on the site, the bot can alert human operators when to offer help or lead visitors to initiate a conversation themselves. For example, the bot flags customers who have errors or can’t get to the next step in a form to alert an operator to offer assistance via the media the customer prefers: audio, video, text, or phone. After all, it’s often impossible for companies to engage, or even view, every single customer at once with large websites. Overseer offers the omniscience needed to let businesses deliver a personalized customer experience even with a massive volume of traffic.
However, Overseer doesn’t convert customers on its own. The people behind the site take the customer to the next step. The bot flags Customers get a special VIP designation for customers passing along personal data, and Overseer reminds them they have access to live help at any time. The bot makes sure that the highest-value customers are more likely to convert by prioritizing them for personalized help if they should request it. The bot also looks at the number of operators available. It won’t trigger an engagement if everyone is occupied.
How We Got Here
Unlike similar bots, Overseer looks at both those customers who visit the site and those who run the site in real time to the second. It factors not only visitors’ engagement but their history with the site. Glia’s bot is the product of 50 years of evolution, starting with ELIZA, the first chatbot that mimicked a psychiatrist’s active listening in response to text input.
Back then, people “talked” with bots back and forth via text. Now, bots can communicate through any combination of text, pictures or actions, sometimes through machine learning or natural language processing. Overseer still has room to grow. Keeping track of millions of visitors online and tracking their every action alongside thousands of operators in real time with no time gap is difficult to do.
Into the Future
Overseer’s future is all about getting smarter. Visitors have probably fed the bot as much data as it will ever see. However, how Overseer interprets that input to create a positive output can improve in terms of machine learning. This means that the bot continues to learn and tweak those results automatically.
When people first hear about Overseer, they often wonder how much input and output the bot can take. They also think about how its business logic could help their business, and how much data they need to give to the bot in order for it to help them.
Want to learn more? Request a Glia demo to see Overseer in action.