How to remove negative associations with bot-based interactions

Chatbots can help humanize AI and increase customer engagement, says LogMeIn director of customer engagement technologies Ryan Lester.

How to remove negative associations with bot-based interactions

TechRepublic's Dan Patterson spoke with LogMeIn's Ryan Lester and discussed how companies can combat the negative associations customers have with AI and chatbots. The following is an edited transcript of the interview.

Dan Patterson: Artificial intelligence and chat bots are dominating how we interact with machines, with websites, with mobile apps, and with IoT, but often there is an uncanny valley of perception. For Tech Republican and ZDNet, I'm Dan Patterson with Ryan Lester, LogMeIn's director of customer engagement technologies. Ryan, how we remove some of the negative stigma or bad taste that AI chat bots can leave in our mouth?

Ryan Lester: I think of a lot of this baggage comes from the fact of this technology's been around, at least virtual assistance has been around for a number of years and historically they have been very scripted. Bots were told, if you get this keyword gives this response. They really weren't built to be conversational and customers interacted with them, and they could do one or two things, but as soon as a customer asked a different question, or phrased it in a different way the bot would break. What's new is this technology's really evolved and become much more conversational, so when we interact, just like you and I are where we'll use different words to talk that have different meaning, but the bot understands those, and then can deliver a better outcome to the customer.

Dan Patterson: What is that better outcome? What are the goals that AI chat bots try to achieve?

Ryan Lester: We talk about this as brand defined and customer driven. What that means is you define what you want the bot to do, so do you want to the bot to manage your returns? Do you want the bot to manage, let's say, change my seat on a flight? You're setting up those outcomes and then the bot drives towards those outcomes, it interacts with the customer to get what it needs. There's really a variety of use cases. It can be financial services, it can be related to travel and hospitality, it can be related to retail, so they really have become quite complex and they can do a variety of things. Really, that comes out of the brand. What do you want it to do, and then build the bot towards that use case.

Dan Patterson: How... or rather what are the technologies behind AI chat bots? I know we use artificial intelligence as the catchall phrase, but can help us understand the tech that is driving AI chat bots?

Ryan Lester: Yeah, great question. We get this one a lot. There's really three main aspects of this. The first is the natural language processing, or natural language understanding, which is what is the person saying? What is their intent? That's really around what do the words mean, the sentence structure, et cetera. The second is really around smarter knowledge management, so once you understand the customer's intent how do you then better organize your information to drive them towards a better outcome whether that's an article, a form, a process, or even an agent? Then the third is around this knowledge graph, so our entity graphing. How do you connect an intent with an outcome, so that the bot, over time, learns that oh, when I'm getting this kind of question the customer really means this, and therefore I should drive them to that outcome. Those three key things have advanced over the last two years: the natural language processing, the knowledge management, and then that relationship graph between the intent and outcome.

Dan Patterson: We usually think about AI bots as a consumer tool and the consumer experience, but what is the B2B relationship or the business/technology relationship with AI chat bots?

SEE: These 10 industries are most impacted by malicious bot traffic (TechRepublic)

Ryan Lester: There's two parts of that, we get that question a lot. The first, we like to use this term harmony where we're bringing harmony between the bot and the internal employee. Oftentimes people think about AI as custom use case, so I'm an agent, I'm talking with a customer on the phone or chatting with them, and they ask me a question, the bot can be there assisting me, and giving me suggestions, alternate articles. Basically, the bot is my assistant there to help me with my day-to-day job, so that's one use case internally.

The second is also internal support use cases. I want a bot to handle things around IT questions or HR questions. We all can relate to this, "Hey, what's our policy around vacations or paternity leave," so bots can also be used for internally facing support questions not just customer facing support.

Dan Patterson: What does the future look like? I know when we talk about AI sometimes, especially right now, we think about "Westworld" and the crazy outskirts of technology, but really what does the short term, the next 18 to 36 months, look like in terms of business technology's use of AI powered chat bots?

Ryan Lester: There's two important things we see. One, we actually see us driving a lot of value for internal employees, so bots can start to handle much of the mundane, those commonly, frequently asked questions that having a live person support is really not a great job, it's not a lot of fun. A lot of that mundane, highly repetitive tasks can be all sent to the bot, so you can have your employees spend time on more high value interactions, so that will continue to happen. We don't see us replacing employees, but rather having your agents and employees spend time on more valuable activities whether that's customer acquisition, or higher-value customers. One is bots will continue to manage the mundane.

The second is bots will become more specialized, so you'll have bots that will become better at things like warranty related items, or related to re-booking a flight. Instead of a bot being an assistant you could almost think about it as becoming specialized to certain skills because it'll get better the language processing related to that skill and it'll build better knowledge related to that skill. You'll start to see bots really become experts in certain activities as they continue to grow more knowledge and capture more data.

Also See

  • Special report: Cybersecurity in an IoT and mobile world (free PDF) (TechRepublic)
  • Humans, cover your mouths: Lip reading bots in the wild (ZDNet)
  • How to tell if your website has been visited by a bot: 5 tips (TechRepublic)
  • Twitter has a spam bot problem -- and it's getting worse (ZDNet)
  • Beware of the bots: How they're created and why they matter (TechRepublic)
  • 20180426ryanbotsdan-1.jpg