Before most of us knew what a bot was, Dan Miller, lead analyst and founder of Opus Research, was already looking at the potential impact the nascent technology would have on customer service and engagement. Now that bots, intelligent digital assistants and other conversational interfaces are becoming daily parts of all our lives, businesses of all sizes are expected to invest close to $5 billion by 2021, according to a new report from Opus Research – Decision Makers’ Guide to Enterprise Intelligent Assistants; The Brave New World of Bots and Virtual Agents .
Dan shares some of the key findings from the report, including what is driving the current bot/intelligent assistant craze, how businesses are approaching things, the role of “metabots”, and the role artificial intelligence/machine learning is playing (and will play) in bot adoption.
Below is an edited transcript of our conversation. Click on the embedded player below to hear the full conversation.
* * * * *
Small Business Trends: So tell us about this report.
Dan Miller: it’s our second look at what’s going on with what we call Enterprise Intelligent Assistants. Everybody’s got their own term for what a bot is, but we were calling it intelligent assistants, or intelligent virtual assistants, and it’s the combination of technologies that fall into two domains. One, I’ll just call the smart UI or conversational front end, that lets people input information, questions, whatever through text, through speaking, which then gets transcribed as text. We’re seeing emotion detection in that category, and we’re seeing biometric coming into that as well, because if you have strong confidence that an individual that’s talking, is who he or she claims to be, you can then make a more individualized and personalized experience.
So, that’s what goes on in the front end. There’s a back end, which is often associated with contact centers, or e-commerce websites, and that sort of thing… It’s natural language understanding. So, people can use their own words to accomplish things in a self-service mode.
We observed, three years ago, and answers “why is this all happening now?” Is around that ability to just let people talk to stuff, or type stuff in their own words, and accomplish what they’re trying to get done. Net-net, it’s made self-service less of a dirty word. We used to think of self-service as people reaching … Or companies basically just trying to save cost by automating as much of the interactions they have with customers through their support resources.
Now, we’re finding a growing number of people really want to self-serve. It’s more like an electronic game now, and hence, here we are. You dial in, use your own words, and you take command with them. That’s why it’s so interesting right now.
Small Business Trends: You talk about two broad technological domains that encompass this whole area. Conversational technologies and intelligent assistant technologies.
Dan Miller: I spoke a bunch about the conversational front end, so let’s move to the intelligent assistant technologies. I just know that people need to map these terms into the words that they use. We see it covered in a lot of magazines as artificial intelligence. Meaning, that there’s in addition to this natural language understanding, which is you use your own words, there’s knowledge management, which is the organization of all that big data that we read about all the time. There’s intelligent search that takes place. There’s analytics, the categorization of all that information that companies have.
So it’s really relevant for large enterprise, who have made investment in CRM with the voluminous records about their customers. We’re into knowledge management. It’s sort of an all-encompassing term, but it’s data management stuff. That’s where your marketware is, where your product descriptions, where training materials, where marketing and advertising are. That’s all machine readable now, and in this dream world of artificial intelligence, the ability to find the right answer, and establish a truth to provide to customers, can be done largely in automated fashion. And where that used to support how an agent answers a question in the contact center, be it screen pops and stuff like that. Or, how a speech-enabled IVR would provide correct answers over the phone in an automated environment. That’s all the resources that we call AI in the enterprise.
The answer, it’s sort of like Dorothy in Oz, right? We’re largely already there. We’ve made investments that are very close to moving to real-time responses to our customers, through these automated needs.
Small Business Trends: One thing that is obvious when reading the report, and there’s a lot of great things in this report, but the investments being spent in the forecast you guys put together. It really looked like it started almost from next to nothing, generally speaking, in 2015, to now you’re projecting 4 and a half to 5 billion in spending on enterprise intelligent assistants by 2021.
Dan Miller: Right, and what’s in that investment does encompass a lot of those technologies that I just talked about, the knowledge management, the natural language understanding. We see that there’s 20 or so vendors with intelligent assistant platforms. Some are household names. IBM would be there with Watson, Nuance, going up with Nina. I don’t want to offend anybody, but trying to rattle these off and then forgetting one or two. Some of what goes into our forecast is investment in the existing resources that are providing good answers to people contacting a company.
Then, some of it is packaged by firms with the platforms for developing virtual agents, and that becomes a very interesting way of looking at the world. That, if you’re in customer care, how do I hire one of these virtual intelligent assistants? That’s why we used the report to list the 28, or so companies and give the criteria for having a look at them.
Small Business Trends: How does what’s going on, on the enterprise side, when it comes to the intelligent assistants and conversational technology – How does that compare and contrast with what the more larger, general audience of consumers is seeing when it comes to, they’re using it on Alexa, or on Siri and things of that nature?
Dan Miller: Right, and we have a term for those unbranded, free-floating intelligent assistants. It could be on your smart phone, it could be on your television. We’ve heard it could be on your microwave, and all that sort of stuff. We call them Metabots. There’s two ways to look at this. One is, as a consumer, you’re going to encounter more devices that have microphones and other sensors that can respond to you, and in some cases anticipate what you’re asking for. You’ll notice Google Assistant has access to your calendar, probably reads your email if you’re using Gmail. It knows a lot about what your intent is, and can support that.
So, these free-floating metabots are the sources of the first point of contact for reaching the companies that you indeed want to do business with. So, that’s the consumer perspective. Now, for companies that invest in their brand, in their own marketing efforts, in ways to reach these customers, in the world of intelligent assistants, they need to regard each of these emerging metabots basically as a channel, and try to minimize the amount of customization, or what I call is exposure to stranded investment, that comes if there’s a lot of effort that has to be made to conform to Alexa’s API, or to make a Google skill, a skill for Google assistant, and that sort of stuff.
So, my long-winded way of saying, whether you’re a large company, or a small business trying to figure out how to reach people through these general purpose metabots, I feel like we’re reliving the world of mobile apps all over again. Or making a decision about how you end up driving traffic to your business in a world where there’s so many alternatives. So, you just regard these metabots as,’ hey it’s just one more channel’, one more digital channel that I should exploit in order to drive feet to my store, or do business online and that sort of stuff.
Small Business Trends: It’s funny, because I had a conversation, this has to be probably close to two years ago, with Dennis Mortensen over at x.ai. With Amy, calendar scheduling assistant, and I asked him “do you feel that once the big guys get onboard, the Alexas, the Cortanas, the Siris, is that going to cut you out. You guys that do one thing, and do one thing good?” And he said “no, what I think is going to happen, is those kind of “megabots”, or mega AIs, as he referred to it “they would actually be the ones who would orchestrate the usage of the various bots and assistants that one person will have.” And so, instead of taking over everything, they’ll be the conductor and be able to help an individual leverage bots in a more structured way. Do you see that as something that could possibly happen?
Dan Miller: Oh, absolutely. If you look at what’s going on with Alexa, for instance right now, I can say “Alexa, ask Capital One what my balance is” and …
My friends from the automated speech business are experts in dialogue design, and now we’re trying to figure out how you deal with a “trialogues”, because there’s a person, his or her bot, and then the business you want to talk to.
What we have right now, is an unsteady state, meaning I can do it. I can have Alexa talk to Capital One, I invoke these skills, and one of the first things you’ll see Amazon’s development folks doing, is getting rid of that step. So, the neat thing about these intelligent assistants, is that they’re learning.
Machine learning is a big part of that. Learning in general, is as well. So, we keep worrying about the bots replacing people, or they’re going to be our mechanized overlords, but the way they learn, we teach them. If we can’t teach them to be better at serving what we want to do, then shame on us.
This is part of the One-on-One Interview series with thought leaders. The transcript has been edited for publication. If it's an audio or video interview, click on the embedded player above, or subscribe via iTunes or via Stitcher.