Artificial Intelligence is without a doubt one of the hottest topics in the world of customer experience right now. But with hype comes a considerable amount of fear, as organisations are becoming increasingly worried about getting left behind or failing to properly commence their AI journey.
Verint, a global leader in customer engagement solutions, recently acquired conversational AI (CAI) company Next IT and has been rolling out CAI solutions to its customers. To learn more about CAI and Verint’s “blueprint” for AI adoption, Fifth Quadrant spoke to Jason du Preez, Verint’s VP Sales Engagement Management APAC.
FQ: Why did Verint acquire Next IT?
JdP: Next IT was acquired to help further enhance our customer engagement strategies and as an investment in advanced cognitive technologies such as, machine learning and conversation analysis. NextIT’s “Alme”, an intelligent virtual assistant (IVA) platform that uses CAI to automate machine to human communications, in a manner that emulates real human conversations, similar to your contact centre agent talking or chatting directly with the end customer. The objective is to automate the solving of customer problems, then use AI to identify what is or isn’t working and further refine the customer engagement.
FQ: So do you see a CAI-driven IVA as distinct from chatbots?
JdP: To me, Chatbots sit at the simpler end of the spectrum - they are designed to answer very simplistic and clear-cut questions. At the opposite end of the spectrum, we have IVA’s which are capable of more complex self-service tasks and a high level of natural language processing to create a more authentic interaction.
As an example consider interpreting the intent of two different customers asking to update their comprehensive car insurance policy, one goes on to say that she just got married, the other indicates she is changing address. Chatbots fail to pick-up on these intonations at this point whereas CAI interprets this information in context to the conversation in a manner that a human having the conversation would.
FQ: We often talk about the concept of “handover” as the moment when an IVA moves a customer to a human agent. Do you see this handover as occurring later as IVA’s improve?
JdP: We are definitely seeing AI doing increasingly sophisticated processes. For example, Dell’s AI engine takes customers through the whole purchase process of buying a new computer, but instead of talking to you about specs it will ask you about use cases. Will you use it for gaming or do a lot of design work? The AI can then build up a persona and make recommendations based on your needs. Having said that, I think there’ll always be a place for humans though.
FQ: How does CAI improve CX and EX?
JdP: Human-like conversations can really build trust between businesses and their customers. Recently, Gartner talked about how organisations implementing IVA's have seen a 70% reduction in call, chat and email enquiries while at the same time seeing significant improvements in customer satisfaction.
Employees can see the efficiencies that IVAs bring as they can automate a lot of tasks for them. We have implemented an IVA named “Julie” for rail company Amtrak, for example. Julie helps customers plan journeys across the country, picking the right options and methods for each trip.
What we’ve also deployed internally something we call “Super Julie”, an IVA that supports agents working in Amtrak’s contact centres. Instead of worrying about booking complexities, agents can use their softer skills to engage with customers better. Amtrak saw an increase of about $1.15 million per month in sales after implementing this solution. It is the perfect example of a hybrid workforce where employees are working hand in glove with AI to create a better outcome for customers.
FQ: Next IT has developed CAI for Alaska Airlines, Amtrak, Merrill Lynch, Teva Pharmaceuticals, and the U.S. Army. Since Verint has acquired Next IT, are there any particular verticals you have been focusing on?
Verint’s typical customers are high-volume business-to-consumer organisations or government organisations that have complex problems to solve. The acquisition of Next IT fits very well with our customer base and the problems we solve for them.
FQ: And are you currently working with any Australian organisations? How is Australia faring when it comes to AI adoption?
JdP: We are working with one of Australia’s largest insurance companies and talking to numerous other organisations. We are finding that most of the larger Australian businesses and organisations have already embarked on some level of AI journey which I think is going to drive greater acceptance of AI.
FQ: Can you tell us more about this insurer?
JdP: The insurance company was failing with their previous AI programme, specifically as they were having problems scaling it. Using our AI Blueprint that we launched in April, we were able to help them understand how to be strategic about their use of AI and get measurable results.
We took unstructured data such as call and chat transcripts from the contact centre as well as search results from the insurer’s website and fed them into our CAI engine. The engine identified five top areas that we could focus on to deflect 72% of traffic back into the call centre. That recently went live after a four-month implementation period and today it is responding to 98% of questions that are coming through. As a result of this success, the insurer is looking at implementing more AI solutions across the organisation.
FQ: What are the biggest challenges to an organisation adopting AI?
JdP: We think the biggest challenge is for organisations to first identify where they need to start and this is something that our Blueprint really helps with.
The other challenges are around defining the problems AI can solve and justifying the business case. The Blueprint approach allows us to pinpoint specific areas to apply the AI to and ensure that we are focused [on] real business outcomes from the beginning. We see it as an agile approach – it is short and sharp sprints quickly delivering value to businesses.
FQ: So how does the Blueprint work?
JdP: What we’ve done is focused on five specific use cases in which there is real ROI to be achieved via the use of IVAs. There is a “Customer Service Assistant”, a “Brand Engagement and Loyalty Assistant” and an “eCommerce Assistant”. The remaining two are slightly different and more internal: an “Employee Resource Assistant” for IT help desk support and a “Human Resources Assistant”. These can help larger organisations in particular with internal processes and support EX.
FQ: What do you think about the concept of “bot disclosure”? Do you think customers should be informed that they are dealing with a bot?
JdP: Absolutely, I think that bots should be disclosed. As chatbots and IVAs become more mainstream we are going to see them everywhere. If you’re not upfront with customers they are going feel misled and it will have a negative affect on your brand.
FQ: I agree. But they should still be human-like.
JdP: Yes, we have given a number of the IVAs we have developed a personality. For example, Sgt. Star, our IVA that does the job of hundreds of recruiters every day has got a real personality so if you ask him about his home life, he will tell you that he is married and his wife is the boss. Giving them a personality with elements of your brand is quite important I think.
FQ: Where do you see CAI in five years? Will it be at a truly human level by then?
I think five years is a bit of short timeframe for that. It is moving at a very rapid rate and there are a number of businesses coming up with bots that serve specific purposes. At the same time, other players such as Google and Amazon are coming out with AI platforms. Off the back of these platforms, I think we’ll see wider adoption and even faster progress.
Will AI be answering all the questions? I think at the end of the day there will always be that need for human intervention at some point.
This interview was edited and condensed for clarity and length.
About Verint Systems Inc
Verint® (Nasdaq: VRNT) is a global leader in Actionable Intelligence® solutions with a focus on customer engagement optimisation and cyber intelligence. Today, over 10,000 organisations in more than 180 countries—including over 85 percent of the Fortune 100—count on intelligence from Verint solutions to make more informed, effective and timely decisions. Learn more about how Verint is creating A Smarter World with Actionable Intelligence® at www.verint.com.