When a restaurant is thankfully busy, phone etiquette tends to suffer if a caller can get through at all. And it’s especially bad at small businesses and franchise operations. 

That was the foundational issue for CallJoy general manger Bob Summers. He came up with CallJoy while he was working on local ads at Google but as advertising drove consumers to a retailer, they frequently got bad phone service. 

“As a consumer, I found that frustrating. When calling my favorite restaurants I’d frequently be put on hold,” said Summers. “Sometimes I’d have simple questions like, ‘Can I take my dog?’ I just want to find out if they’re dog friendly and couldn’t find that online.” 

So he started working on the issue during what Google calls 20 percent time, unstructured project time during which Google employees can work on their passion projects. He and another Google employee experimented with some solutions and it seemed to work. So the project found a home within Google’s Area 120, the company’s internal springboard for those passion projects. 

“What Area 120 does is allow you to do that full time, it’s essentially a startup incubator,” said Summers. 

As the project progressed, he said the team found a lot of fascinating insights into the gaps small business owners have with their phone infrastructure. Unlike the big companies that Google works closely which tap into all sorts of data or route customers through a call center, for small businesses (as most franchisees know) it’s just a phone.

“There are two aspects of CallJoy that are important to understand, first, call analysis: what is happening with my phone. This was a big surprise when we started, we found that small business owners didn’t have any insight into what was happening with their local phone,” said Summers. “They couldn’t see how many people called, how long people stayed on the line how long they stayed on hold. Basically they didn’t have any data around call quality.” 

To address that aspect, he said they sought to measure phone usage as a starting point. But even that was arduous. 

“We asked them to look at their call records to see how many calls you get per a month. It turns out with landlines, you don’t get a list of calls. In fact, to get those records, you have to use a court ordered subpoena. AT&T literally said they would have to use a court order because of federal regulations.” 

That prompted Summers and team to just plug those calls into the world of Google where it could be recorded, logged and transcribed. Ultimately, the idea was to classify those calls between reservations, questions like parking and food allergies—basically determining the intent of the calls. 

As they did, business owners learned all sorts of things like their training was wrong and folks manning the phone were giving bad information and how to measure marketing programs that prompted calls. 

“Then the second step is automation,” said Summers. “Is there a group of things that we could be assistive to that team or customer service team and provide an immediate response?” 

He said that second aspect helped the system “scale infinitely for the consumer” so that no matter when a call came in, during a rush, after hours or when the only person who knew the dog policy was on break, the system could answer those key questions on the phone or via text messages. 

“We’ve brought Google AI essentially to the small business owner,” said Summers. “The tech we’re using is typically reserved for Fortune 500 companies with tens of millions of dollars in budget to build automated call services.”

The newest update that came in November made that AI phone agent a little smarter. Now, it asks the simple question, “How can I help you?” 

“The response to that prompt is the consumer intent,” said Summers. “They want one of two things, they either want information such as do you have Wi-Fi or vegan options are dogs allowed, do you have a band on Friday nights. Or they’re seeking to do some type of transaction like table reservation or a food order or I want to talk to someone in catering—catering is a common intent for these businesses.” 

The update also added functionality like having the AI pick up after hours or a certain number of rings. It also added phone transcriptions that owners can use to find more key questions for the AI or train staff. Summers said it takes about two weeks of monitoring to determine what key questions come up, from there the restaurant owner or retailer can decide how to handle them. 


Please enter your comment!
Please enter your name here