At 6’ 9”, Phil Gordon, the CEO of Chatbox, stands out in any crowd. It doesn’t hurt that Gordon, a former coder who made bank when his company Netsys Technologies was acquired by Cisco, became one of the poker world’s top talents, at the table and as ESPN top poker analyst.
Back-end Hacks for Better Chat
The interview with
Phil Gordon at Chatbox
Along the way, Gordon started investing in startups, but when he stumbled on the idea of a messaging platform that would bring automated, personalized text messaging at scale to enterprises, he placed his bet on Chatbox, which has now become one of the leading integrated messaging platforms, bringing customer experience to new Gordonish heights.
This fall, just as Chatbox concluded a deal with Oracle to bring its new Instant Apps technology — a plug and play development platform to integrate photos, signatures, and other non-text data — into AI-powered Oracle Intelligent Chatbots, I sat down with Gordon and Marco Lafrentz, tyntec’s director of CPaaS solutions, to discuss how new technologies that connect to the evolving cloud communications ecosystem are helping innovators like Gordon change the future of customer experience. And that’s a very big idea.
Q: Did Chatbox start with B2B or B2C use cases?
Phil: Our very first installations was B2B, a company here in Seattle called Coast Workplace Solutions. Coast serves as an intermediary to help businesses service their various locations. If you’ve got a scheduled A/C repair or you need a carpet replaced, you’ll call Coast and Coast will find the contractors to come out and fix it.
But one of the problems Coast had was that they’ve got thousands of contractors around the country, and they couldn’t get them to open email, so they turned to Chatbox for a messaging solution that would enable a Coast rep to text message those contractors for bids.
They actually text a work order to the contractors, get pictures when the contractor gets to the location, get the signature of the business when they leave, and pictures on the way out. Everything’s time stamped and all of those messages and the various picture data are integrated seamlessly into their back end workflows.
Q: That’s an obvious use case, and one I’m sure we’ll see grow. Marco, do you see B2B or B2C use cases gaining the most traction?
Marco: No matter what you do, platforms such as Facebook, LinkedIn, Twitter, are gaining more and more traction, and they are all based on short text comments, so it is quite natural that this has created growing demand to adopt chat for professional collaboration and commercial messaging.
Technically the delivery of the messages is the same as in social media. But when it comes to business or commercial messaging, it’s different. When I text a friend, everybody is aware of the context. In the enterprise, the user experience needs to be more thoughtfully designed. You need contextual information that pulls data from other systems into a communication platform that establishes the contextual relationship of people and channels.
Q: Exactly. Take a travel use case, where I’m on my way to the airport and suddenly get a text from my airline that my plane is going to be delayed. Now the entire context of the flight delay is tied to re-booking, and requires a great deal of back end sophistication. Have you dealt a lot with B2C use cases in that kind of air travel context, Phil?
Phil: We actually have one right now, and one of the interesting things about this particular travel use case is that of the 160 or so global airlines, almost all them offer notifications — your flight is delayed or on time or leaving from gate A23 — but only a handful offer the chance to reply. I’m running late, can I get the next flight? I need to change to tomorrow. I’d like a window seat. Can you please check me in?
Those bidirectional offerings have not been successfully implemented, and I think one of the reasons is that enterprise companies, are having a very difficult time finding solutions that easily connect with their legacy back end systems, their CRMs and databases. Getting that data into a message-able format, and then responding to inbound messages and doing the right thing in your legacy back end is a pretty tough problem for them to solve.
We’ve been able to crack that nut a little bit here, and we’re going to be helping that global partner with that exact use case. So not only will notifications go out on those channels, but we’ll also be able to facilitate flight changes, rescheduling, seat selection, and all sorts of new use cases.
Q: What other verticals are investing in that kind of bidirectional chat? Who’s successful?
Phil: One use case that stands out is helping companies actually facilitate phone calls. Think about this use case: an educational services company selling high priced classes in complex sales. You go to a website, fill out a form, hit submit, I’m interested: Even at the best companies, it’s going to take two or three minutes to get a call. But by that time, no one’s picking up the phone.
Even if you’ve expressed interest in receiving a phone call from the company, 95% of those phone calls go unanswered. But one of the things we’re seeing is that if you can text first and say “hi, this is company ABC, I’d like to talk to you about the product you just inquired about. When’s a good time for a call?” you get unbelievably good response rates. We’ve got one company that increased bookings by 30% with that solution.
Marco: In fact that’s one of the best things about texting: it’s totally non-intrusive. Sending somebody a text message doesn’t need to be context aware. You can be in the bedroom, on the toilet, driving, on the run — the text isn’t interrupting you.
Phil: It’s actually more respectful of my time than a voice call. Say I’m in the middle of a meeting and I don’t want my phone to ring. It’s coming from an unknown number. But if you can make it clear that that text message is from the company — “Hey, it’s company X, when’s a good time for a call?” and have the phone ring from that exact number that you texted from — wow, you’ve really closed the loop. Now, it’s not an unknown number. I just responded to your text message. I know who you are. We’ve seen some early signs that connection rates are more than 70% higher than they were getting before with just a phone call.
Marco: Voice is less trusted than ever. There’s just so much fraud in voice that people don’t want to pick up the phone anymore. The text channel is protected. People trust texts. You open them, read your messages. Abuse is extremely low. And this is due to the technical differences in the set up of the back end, the way commercial and personal messaging have been separated. Text is the most trustworthy channel and probably why it even increases the trustworthiness of communications in the voice channel when they are integrated.
Q: What do you see as the strongest component in the Chatbox tool kit?
Phil: One of the most important things is automation. When a user submits a form, you want a text message to flow automatically so that the back end response seems natural. So we might send a message such as “What time is convenient for a call?” then use natural language processing to discover the right time. If they say “now,” you route them to an agent right away. If they say “two hours,” then you route the agent to them in two hours. That whole experience can now be managed and automated as much as possible up front with back end data so the messages are deeply personalized.
If you say, “Hi, Phil. This is Tom from company ABC. I know you’re interested in product ABC. When’s a good time for a call?” you get much better results than if you say, “This is company ABC. When’s a good time for a call?” It’s that personalized experience that requires the integration with the back end systems that transforms the conversation and creates that more personalized feel for the end user.
Q: These are principally decision trees within a machine learning context, correct?
Phil: Well you really do need some natural language processing there in order to operate at scale. Because someone’s going to say, well how about after lunch? Or, I’m available in a few hours. But, of course, natural language processing isn’t as good as people think it is. Take that simple question, “what time should we call?” I’ve run some tests against Google’s API and Amazon’s Alexa, and they’re getting about 60% of those responses right. Now, you can get higher if you are willing to do more work there. In our solution, we can get about 95% of those responses to the correct timestamp.
But if you’re going to run this a million times a month, even a 5% failure rate gives you 50,000 failures that you’re going to have to escalate for a human response which is why Chatbox provides an omnichannel messaging solution including automation and the natural language, the agent tool for escalation, as well as the integration with your back end systems to hook up all the data that you are receiving, so if someone says can you contact me at 12:15, we can actually create a sales case for you, sync it to your calendar, or issue a call back to your APIs to sync to your call center solution.
Marco: If you really want to solve the integration problem for the enterprise, you have to respect that their infrastructure is also deeply fragmented. There isn’t a centralized solution for their IT or for all the touchpoints they would like to have. There are three components to this: First you have to play the game for integrations and provide them with APIs. Second, you need to provide for integrated use cases into CRM systems, predefined frameworks and surroundings. Lastly, you need to give them an opportunity for measurable deployment, a live dashboard with a direct graphical interface.
This conversion into the API layer — routing voice calls, sending text messages, integrating all channels, getting number verification services — is what really powers enterprise solutions, It’s the key change that allowed them to rethink how they handle communications and customer experience.
Phil: Companies like yours, Marco, along with Twilio and Zipwhip, have democratized access to these really powerful concepts. We build platforms on top of solutions like yours — we’re basically the value added platform on top of the messaging solutions. Then the customers can build their solutions on top of our solution. Without that underlying API access, none of this was even possible a few years ago.
Q: If you had a crystal ball, how would you say B2C businesses are going to shift as a result of these value added solutions?
Phil: 2018 is going to be a very important year for enterprises in messaging. The Facebook Messenger bot platform was only announced 15 months ago, and kicked off a frenzy of activity around messaging based solutions, bots, and Chatbox. I think the budgets are going to get unlocked for these solutions in 2018 and you’re going to see these messaging based solutions as a true differentiator for companies that are able to get their act together enough to the build solutions for their customers.
Their customers want it. Nine out of ten U.S. customers report wanting to message with businesses. U.S. adults spend 23 hours a week in messaging solutions, right? They’re already there. They want to interact with businesses on these channels. And the businesses that are quick to offer those solutions to companies and to customers are going to be the big winners.