Inference CEO on the Evolution and Future of IVR
In this Executive Insights interview, Blair Pleasant speaks with Callan Schebella, CEO of Inference Solutions. Schebella discusses the Evolution of IVR – where it’s been, where it’s going, and how the cloud plays an essential role in this evolution. He also talks about how the integration with apps and event-driven workflows, as well as AI and machine learning, will change the way we think of IVR.
Blair Pleasant: Hi, this is Blair pleasant and I'm with Callan Schebella, CEO of Inference Solutions. Inference offers Inference Studio, a cloud-based platform for telecom service providers; it gives them tools to design and build IVR systems, to offer customer self-service, call automation, and network routing over a telephone and messaging channels. Welcome, Callan.
Callan Schebella: Hello. Thanks for having me.
Blair: Callan, IVR has essentially been around since the 1960s, and it started being used in most contact centers in the 1980s and here we are still using it today. Tell us about how you view IVR and how it's involved over the past few years.
Callan: Yes, you're absolutely right – it’s been around for a very long time and I think in a lot of ways it became stale, particularly in the 80s, 90s and 2000s, where people were looking to alternative technologies. One of the reasons why it did become stale was I guess part of its history, in that IVR systems when they came out were expensive platforms, they were complex platforms, and in a lot of industries if you have something that’s expensive and complex people tend not to want to touch it once they’ve got it up and running.
I guess over the last few years we've seen a couple of big developments in the industry and I think probably the biggest one of those would be the availability of high quality public cloud networks and private cloud networks. Particularly those that can handle real time voice communications. If we just go back a few years there wouldn’t be many people that would trust a public cloud platform for their real time voice data, but that has definitely changed over the last few years. And then I think coupled with that is that we have big developments in the explosion of cloud technologies, as they’re available, so self-style technologies that you can actually reach out to and consume and therefore improve IVR and IVR-like services.
The types of services I'm talking about here are things like speech recognition on demand, text-to-speech on demand, voice biometrics, speech transcription, speech translation – lots of really interesting and cool technologies that you can now basically consume in an on-demand like manner. If you put those two things together, what you've got is a much more sophisticated range of technologies that you can draw from, but also at a much lower price point. And if we can just compare the systems we have today to those from 10 or 20 years ago, we are talking about an order of magnitude reduction in price and that also manifests itself in the amount of time required to sort of put these systems out in front of customers. I think those would be the biggest trends I’ve seen in the last few years.
Blair: Okay, and that makes sense but let me ask you, how relevant is IVR when everyone is using either the web or mobile devices to interact with companies?
Callan: Yes, the role of IVR has changed over time. Inference, as a company, for example, has changed with that. We’re primarily known as an IVR company, that’s sort of our heritage, that’s where we came from. But if you were to kind of look at the services that our customers build on our platform today, and I think if you were to ask most people how they would describe that service, IVR is not one of those terms that would come to mind.
By that I mean things like proactive notification-based systems, or proactive authentication, determining who the caller is based on the way they’re speaking. SMS messaging services, chatbot services are not traditionally seen as IVR, but they are interactive self-service response systems.
And then another area where I think that the role of IVR is blurring, is what I've called event-driven work flows and in many cases these are events driven in many cases by mobile applications for example, an application on a smartphone. Whilst the mixture of channels that people choose to use has changed dramatically over the last 10 or 20 years, the telephone channel, the voice channel, is still super important. It might not be the first choice that somebody's going to today, but it is still the choice that people will often turn to when something is not going correctly or something is very important to get right that we use a voice channel. So I think it is the mixture of channels that has really changed over the last 10-20 years.
Blair: Yeah, I'm glad you mentioned the event-driven work flows because that's something that is obviously really important these days and we're seeing more and more demand for that type of thing. So can expand on that a little bit and what you're seeing out there?
Callan: Yeah, absolutely. One of the classic things that people would like to see in their application, or app, is the ability to say, “Hey, just call me.” And maybe I want to send a text message to a contact center to say “call me.” And they don’t want to wait in a queue, they basically want to trigger an event that says, “Hey, this person needs to be called back,” will virtually place them into the queue, will wait for them to work their way to the front of the queue, and then will ring them back and so everyone is happy about that interaction. The agent is happy; they're talking to a customer who hasn't been waiting for a long time. The person initiating the call hasn’t been waiting on hold for a long time. But that’s just one example.
Another thing we see with event-driven reactions is sort of switch-driven events, so things that are happening at the carrier network level. Hey, this queue was getting particularly long, or this part of the network is unreachable, or perhaps when a call comes in let's work out who you talked to you last time and see if they're available. All of these things sort of happen behind the scenes without the caller really knowing, but what it results in is a much better experience for the customer and ultimately for the enterprise that they’re dealing with.
Blair: Ok great. You mentioned cloud and SaaS and obviously we're moving into a cloud world or we’re in a cloud world, but a lot of people feel that IVR and that type of thing should be on-prem for security reasons. Tell me a little bit about where you see IVR fitting into a cloud environment and how things work as far as security compliance and all that.
Callan: Yeah, this is a really interesting question and I think part of it is because of how things have been done in the past or the perceptions around cloud were in the past. And those are really changing at a really fast rate. One of the things I would say here is that just because something is on-prem doesn't mean it is secure. In fact, we see a lot of customers moving to the cloud because the cloud in effect has a higher security burden; it's held to a higher standard because it has to be. And so often the rigor associated with a cloud platform is going to be much higher than a premise-based platform and so a few examples would be when was the last time you did penetration testing on your premise-based IVR platform, as a question for a customer, or when was the last time you reviewed the physical security of that large box in the data room at the end of the hallway? Who can actually get into that? When did you have a third party audit look at your compliance and your security?
Cloud IVR is subject to all of those things on a very regular basis: 24x7 monitoring, proactive intrusion detection. All of these sorts of things that, for many organizations, aren’t really cost-effective to implement. But when you’re operating in a cloud model you’re basically getting the benefits from all of that extra engineering and extra process that a cloud vendor really has to go through to offer these sorts of extra services.
Blair: Okay, that makes a lot of sense.
Callan: I think also one point here is, if you look at an organization, or a pool of organizations, and you said, how many of you were using Salesforce.com, or how many of you are using Microsoft Dynamics, how many are using an outsourced call center or offshore call center? That's where your most sensitive data is being stored, and none of that is on-premise. It's all stored somewhere else.
Blair: Yeah, it sounds like we just have to overcome some perceptions that we have about the cloud. Everyone you see is in Salesforce and having all that important data in the cloud. So obviously it is secure…
Now, as a consumer, most of us have a love-hate relationship with IVR. I know I do and yeah, it's great for saving time, to be able to do self service, but it can be infuriating. What do you see that can be done so that IVR is more user-friendly and less frustrating to consumers?
Callan: At Inference we like to talk about four main things that you should look at when you’re looking at how to make your IVR better. The first of those is to make it personal. I’ll talk more about that in a second. The second would be to make it proactive; make sure it integrates with all your other customer contact channels. And fourth and this one's often overlooked is make sure it’s current; make sure it's up to date.
Going through that list, when we look at personalization, we’re talking about that really the IVR should know about the caller. It should have a good idea about why they’re calling, who they talked to last time, perhaps it knows your name, it's Integrated into your customer management system, it makes it much more contextually relevant to the caller. So if you feel like you are having a personalized experience, your level of engagement is much higher.
And given the availability of these salesforce.com and CRM-type systems, there’s really no excuse to not have a personalized experience. And this can come down to indeed the actual customer contact channels entirely: this customer contact prefers SMS; this one here prefers web chat; this one over here prefers a voice call. It should be personalized like that.
The second category, being proactive, is that we often see people using IVR at two points in their organization: where they solve problems and where they make money. And these are two great points to be involved in. But really putting it as literally a reactive service is a poor way to do it. Making it proactive really changes the way that people think about IVR.
For example, you want to ring in and report an outage. Okay, that's a reactive IVR – “my service is not working;” “my cable is not working;” “my electricity is turned off.” Now let’s turn that to proactive. When that power comes back on or when that service is restored, let's ring you and tell you about it, but not only tell you about it, let's ask you to check and say, is it working for you? “Yeah, my cable modem’s back on.” “Great - we're gonna close that ticket out for you now.” And do that completely automatically. That's a really good experience from a customer perspective, because it keeps him in the loop and it's totally proactive.
If you're ringing in and your flight’s delayed, why not tell you that as soon as possible as soon as you get on the call, if I know you’re on a flight. If your prescription’s ready for pick-up, I know it’s a super common one, but I'm going let you know, or change the pharmacy where you pick it up for example or whatever. Making IVR proactive really changes the relationship that it has with the caller.
The third one would be integrating with customer channels, and this is the classic IVR frustration, where you ring into a system and it says, “hey, please give me your customer number,” and you go through a menu tree and then you go through an agent and they say, “OK, what’s your customer number?” It's like, “I already told you that. I looked that up before and I put the card back in my wallet. I don't want to have to do that again.” And this becomes even more important as you're moving between channels with a single interaction. They might be having an SMS interaction that escalates to a voice call. All that information should be preserved. You may have a web chat that escalates to a voice call. All of that information should be preserved. So I think that's another area where we see things falling down and you have more of a hate relationship with IVR.
And then the last one is keeping things current. Keep it in sync with your business, effectively. If you bring out a new product, people are likely going to want to talk about that products when they call you, or they text you, or they chat with you. And a lot of the reasons why you ring into organizations where the IVR is not current is because IVR, and it goes back to that point I was making earlier, was complicated, it was a bit of a black box, no one wanted to touch it in case something broke, because they didn’t know how to fix it… and it just sort of got left there to just sort of live out its life in a sub-optimal way. And I think that’s another big area where IVR falls down.
And I think finally, to add a fifth one, one of the areas where IVRs can add value is in the analytics side of IVR. IVR platforms, by their nature, can capture enormous amounts of data. But are you using that data? Are you linking that IVR into an analytics platform to work out which paths people are taking, when they take them, and why they take them, to offer a better experience? Whether it be a customized experience, or a completely different self-service approach. That would be another area to focus on as well.
Blair: Yeah, I I love what you said about making it personal and IVR knowing about the caller. We talk about agents knowing who someone is when they call, but it's important that the IVR system also knows that so it can present the right menu at the right time so that people go through it a little more easily and much more intuitively. Agreed.
Callan, generally it's been complex and hard for companies to develop IVR applications. What's different today?
Callan: Well, at Inference a lot is different today. Our view is that if an IVR is easy to maintain and an IVR is easy to build, the likelihood of you getting a better one is much, much higher. One of the things that has changed obviously in the last few years is you have better browser technologies, better web frameworks that you can use, and this allows you to basically offer really interesting, I’ll call it “design interfaces” to customers to use or to enterprises to use.
Our flagship product is Inference Studio. If listeners aren’t familiar with that it's a hundred percent sort of graphical environment where you can build pretty much any sort of a self-service interaction, whether it be on the voice side of things or the text message side of things, like an SMS-based chatbot for example. But the important thing here is that you can do it without technical knowledge. So you don't need to understand how APIs work, you don't need to understand how speech recognition works other than at a very superficial level. You don't need to understand how something like biometrics works to be able to use it.
We see a lot of press about companies like Twilio, for example, which has really revolutionized the communication platform-as-a-service industry. But if you look at it, their kind of market and their kind of customer, they are entirely a technical audience.
For a long time time the tagline was, “ask your developer; they’ll tell you what Twilio is.” Because literally you’re consuming APIs and you’re building applications on top of APIs. Our audience is everybody else; people who are not developers; or people who do not have the time to develop a long and complicated application. They want to get up and running with something very, very quickly. With Studio, you can literally have an idea in your head or a problem that you need to solve, go into Studio, click a few buttons and then be talking to it or texting to it in 90 seconds or less. It's super, super fast to deploy it and then deploy it to any carrier that you choose. So that's dramatically different to how IVR you used to be deployed. Yeah, I think that's what's different today, or at least what we hope is different.
Blair: Speaking of something that's different, you mentioned IVR being more proactive and knowing when someone is calling, what are some other trends you see coming down the road? Like do see AI and machine learning coming into play? What's the future of IVR and where are things headed?
Callan: AI and machine learning are absolutely coming into play. The challenge we had with AI and machine learning in general is they are relatively difficult to understand particularly to a live person. Yes, absolutely they are coming and I think that really coming in two ways.
AI is manifesting itself in the types of dialogue or interactions that you're having in an IVR system. So a move away from what we would call a directed dialogue to a true natural language style application or interaction. And the second one is in the data – analyzing the data that's generated from these platforms and making meaningful decisions based on the data.
If you look at for example, what our customers are doing with our platform, they're consuming things like API AI or IBM Watson, as a few examples of these sort of cloud-based machine learning and IA-driven systems to design sort of natural language engines. So they would just build the engine in API AI, and then they will just consume that engine through Studio and be able to offer over a text channel for example a true natural language chatbot. And the reality is that you can now build one of these things in just a couple of hours without really knowing how the natural language is working at a semantic level. That’s all abstracted away and made really simple to use.
I think another big area that we see is the use of machine running as it relates to things like transcription, where you’re actually getting verbatim all your content and you're converting that into text or even machine translation from one language to another. As well as some more of cutting edge stuff like sentiment analysis and trend prediction is another area where we see IVR being used so there are some really interesting demonstrations out there that you can see online now of sentiment analysis real time observation of two callers talking to each other and how positive or negative they are based on how they sound, but also the words they’re using and the structure of the language that they’re using. So yes, I do see them coming into play and will have one of the biggest affects on IVR as it moves forward.
In terms of where things are headed, I think from an IVR perspective things are going to get more and more blurred, in terms of what an IVR actually is and what it actually means. Whether the term stays relative I guess is the question. The actual technologies will stay relevant, but whether the term still makes sense.
We started as an IVR company, but we do a lot of text processing, we do WebRTC, we do mobile app integration. Are all these things IVR? Not really. They’re interactive, they’re response, but they’re not always voice. And they’re definitely not voice in the form of a traditional telephone call. I think that will see big changes in the way that data is used. I talked earlier about switch analytics. We have invested a lot of time in our R&D, looking at a carrier call switches and being able to make intelligent decisions and reactions based on what's happening at a switch level and this opens up all types of services that we don't see either in the market today like being able to walk into a room and say, “hey, everybody here, we are going become an outbound call center because we have to react to some customer crisis; we’re hitting the phones now,” and everyone's mobile phone turns into an agent. This sort of stuff is really changing the types of services that you could have.
The reality here is that its APIs that are driving all these services. And so your AVI just has to be able to consume these feeds and in fact, eventually modify itself. I think if you look at the nature of the dialogues that are likely to come out over the next few years, you’ll have an IVR interaction that feels more like Siri, and less like a directed dialogue.
And in the same way that Siri might get it wrong you’ll be tolerant with the IVR slightly getting it wrong because you’ll have that sort of relationship with the thing you’re talking to you as opposed to a phone tree. If you pressed 1 and you didn’t get 1, you would be really, really disappointed. But when you talk to Siri and it sort of gets it right and maybe you change the way you say it and then it gets it right it’s a much more tolerant sort of interaction. While I think IVR as a technology or as an area has stood still for a very long time, in the last few years this has definitely fallen away and it is now moving at a very rapid pace and I don't expect it to change.
Blair: Well, and it really does sound like IVR has evolved and come quite a long way and there's a lot to look forward to. We'll have to see what it'll be called in a few years and how things are going to progress so thank you so much for your insights, Callan. It's very much appreciated.
Callan: You’re most welcome. Thanks for having me.