The Business Impact of Chatbots: Two Case Studies

Trends

When you’re scrolling a website, and a chat box pops up — “Hi! How can I help you?” — What do you do?

Some of us close out the window, finding repulsive the very concept of automated customer service. Increasingly, however, consumers are finding it suitable to use a chatbot program as a primary means of interacting with a business. Around a quarter of Americans, according to recent data, are willing to make purchases through a virtual assistant program, and around one-eighth of Americans have already purchased an expensive item through one (and these numbers are only trending upwards). This technology has also broken into other areas of customer service, recruitment, job seeking. Now, you can even order pizza by texting a robot.

The question, for those of us with distinct memories of the stilted chatbots of the 2000s, is this: are today’s chatbots genuinely useful, or are they superficial?

A Chatbot for Sales

A Chatbot for Sales

AllState, the insurance company, made a significant shift in 2016. Targeting small businesses, they began to offer new products, systems, and policy guidelines.

What was useful to businesses was hard work for AllState agents around the country, who now had to get educated on all the new material. Agents began calling the company’s support phone line with a frequency they never had before, to get information on how the new policies worked, and how to get sales quotes.

AllState call lines flooded with urgent questions like these. Those calls backed up. Agents found themselves sat down in meetings, unable to quickly access the information they needed to serve their clients. Therefore, fewer sales were made.

This was a problem. If agents were slow to pick up on the new material, they at least needed a way to more easily and quickly access that which they didn’t know.

The solution was Abbie — or, more accurately, “ABIe,” the AllState Business Insurance Expert. Abbie is a chatbot with an avatar, and a voice, that can understand the context and provide agents with questions and materials in real-time. It is quite complex software:

ABIe finds answers through a combination of contextual knowledge and original content. ABIe relies on component content, taxonomies of key concepts and terms, and curated tagging and search. Content is created to answer specific questions and to provide step by step instructions to particular tasks. Web analytics technology tracks how agents interact with content — what they search for, and what they find, and what is useful. Agents rate the content, and comment on the content and ABIe herself. Data analytics helps ABI identify new FAQs and emerging issues…

Abbie was much faster and more efficient than calling a human customer service representative over the phone. It’s why, three years after release, Abbie was receiving over a hundred thousand inquiries per month from thousands of AllState agents and employees around the country.

Problem solved !

A Chatbot for Recruitment

AllState was dealing with an unprepared workforce, an inefficient communications system, and underperforming sales numbers. An AI chatbot was the key that mostly fixed it all. Hardly an expensive investment, Abbie mostly paid for itself within one year of operation.

Not all chatbot stories work out so perfectly as AllState’s. Still, chatbots provide small conveniences in just about every domain of business.

Take, for example, recruitment. FirstJob, a job placement company, built its own bot — called “Mya” — to streamline the hiring process for hirers and candidates alike. Mostly, Mya takes care of the hiring process between when a job application is completed, and when a hiring decision is reached. It asks supplementary questions of candidates, provides them with updates as progress on their application is made, and connects them with human representatives when appropriate.

To the extent that the content job applications can be converted into raw data, Mya also helps sort out the wheat from the chaff. Its metrics take into account education, experience, and other parameters to rank candidates according to their fitness for the role at hand.

Interestingly, it’s not only the recruiters who have their own bots. The same year FirstJob debuted Mya, Esther Crawford was out of work. So, in her spare time, she developed her own bot. She called it “EstherBot.” EstherBot was capable of interacting with recruiters over Facebook Messenger, offering a selection of presets such as “Childhood,” “Education,” and “Career.” If their interest was piqued, they could hit a button to contact the real Esther directly.

She got eight interviews, and three job offers through EstherBot, and now runs her own company.

Takeaways

Are chatbots useful or superficial?

The answer is that they can be both.

Plenty of chatbots deployed around the internet today are archaic in their design. They pop up in a browser window loudly and disruptively, offering little to those who engage. They mimic human conversation poorly and seek to convert sales rather than help customers.

More often than not, though, businesses use chatbots to good effect. The advanced AI of today is capable of being as or even more useful than human representatives of a business. FirstJob and AllState are evidence of the fact. A more extensive, more robust support line probably would’ve helped AllState with the problems it was facing in the mid-2010s. However, it would’ve been less efficient for agents to have to call every time they had one little question or another, and hiring so many more operators would’ve cost the company orders of magnitude more than Abbie did.

As with any business solution, chatbots can be deployed effectively or poorly. In the right hands, they are a cheap and effective solution to the kinds of low-level problems any business is bound to face. And executives across the spectrum are starting to get the message.

Disclaimer: Being a follower of ‘Chatbots Journal’ offers lots of perks :) A consultation session with experts from across the industries is a major one. Submit your query here and we will connect you with the right Chatbot experts.