IntelliResponse VOICES in Action

Earlier this week I wrote about IntelliResponse’s launch of their VOICES analytical product. I now have some updated screen captures of VOICE in action.  Click on the image of each screen to see an expanded view.

This first bubble screen shows an overview of the common themes that VOICES has picked up from customer conversations with IntelliResponse virtual agents. Based on the size of the bubble, you can see at a glance that debit cards has been a top topic with customers for about the past six months. Customers were also inquiring a lot about mortgage rates, medical insurance, and other related topics during that time.

VOICES 1

Bubble Dashboard with Customer Themes

The second screen shot shows how you can zero in on a specific time slice. During that short time window, some completely new themes popped up. In particular, customers inquired a lot about their account number. This might be an indication that something changed with customer account numbers during that time. It’s certainly something that you’d want to investigate further to find out why the account number topic suddenly became so prevalent.

VOICES 2

Snapshot of Increasing and Decreasing Theme Prevalence

The third screen shows a drill down view of a specific theme. This view lets you learn more about what customers were asking when they made inquiries related to a specific topic–in this case travel rewards. Based on the questions, we discover that customers are interested in credit cards that let them earn travel rewards. This information is gleaned from real customer conversations, so it’s presumably more reliable than data that could be collected from a survey. This would be interesting data for the product development and/or marketing teams.

VOICES 3

Drill Down Into Each Theme for Content

There are many data filtering options available in VOICES that aren’t shown here, but the screen shots give you a good general feel for the tool’s capabilities. There’s undeniable intelligence locked up in the conversations that customers have with virtual agents across all web self-service channels. It will be interesting to observe how IntelliResponse’s merger of virtual agents, customer conversations, and analytics takes hold in the marketplace and is leveraged by organizations.

IntelliResponse VOICES – Virtual Agent Meets Big Data Analytics

Big DataIntelliResponse just announced the official launch of VOICES, a product that merges virtual agent technologies with big data analytics to deliver impressive potential value to businesses. Imagine how valuable it would be if you could read the minds of your customers. What if you worked for a financial institution and found out that many of your customers were starting to think about how to increase their credit limit? Or, better yet, what if you saw these same customers thinking about a new promotion from a competitor that was designed to make them switch credit card providers? Equipped with this mind-reading capability, you could be proactive in meeting customer needs. You could make information on credit limit increases more accessible, while matching or beating your competitor’s promotion. Instead of losing customers, you’d make existing clients so happy that they’d sing your praises.

IntelliResponse’s VOICES product puts this powerful capability into your hands. I was invited to see a live demo of the updated product prior to the official launch. Paul A. Smith, Vice President, Product & Services, walked me through VOICES features. VOICES is an analytical tool that has access to all the conversations that have occurred between your customers and your virtual agents. These conversations can take place in any of the channels where you’ve positioned IntelliResponse web self-service agents; the corporate website, Facebook page, or other consumer-facing channels. The VOICES technology analyzes the unstructured data from these many customer conversations and identifies core themes. The themes are displayed as bubbles in a graphical dashboard. The more often the theme pops up in conversations, the larger the bubble representing the theme. It’s basically the same concept as a tag cloud, where the most common words and phrases show up in the biggest font size. Color coding lets you stay on top of things by showing at a glance which topics are trending with customers and which ones are waning in importance on a given day.

IntelliResponse VOICES

VOICES offers you filtering tools that allow you to look at the themes in different ways. For example, you can filter the data to view only the themes that originate in conversations from a particular source, such as your company Facebook page. What’s more, VOICES allows you to drill down into any theme to take a closer look inside the mind of the customer. Each theme is broken down into subthemes. For example, you might drill down into the theme ‘credit card’ and discover the subtheme ‘how to personalize my credit card.’ If you’re not already offering credit card personalization services, it might be time to think about the possibilities.

The primary purpose of virtual agents, sometimes referred to as web self-service agents or service avatars, is to assist the customer by quickly and effectively answering their questions. Instead of waiting on hold for a human service rep, consumers can engage a virtual agent directly from a web page, a mobile app, or a social media site and get the answer they need. But IntelliResponse has tapped into a whole new source of value with the launch of VOICES. There’s gold in every conversation your customers have with virtual agents. Customers are asking questions about your products and services and giving hints about what they like, what they’re considering buying, or what makes or breaks a sale. Prior to VOICES, this treasure trove of knowledge was going uncaptured. With VOICES, you not only harvest the treasure, but you have tools you can use to make sense of it all.

I don’t have hands on experience with the product. As a cloud-based “software as a service” technology, though, it would seem that an organization could be up and running with IntelliResponse virtual agents and VOICES fairly quickly and easily. I can also see applications for VOICES inside the enterprise. For example, large corporations who run social networking applications within the intranet could benefit from having access to trending employee conversations and themes. The VOICES analytical engine is currently geared towards processing interactions with virtual agents, but I’m guessing that the technology could be easily extended to handle other sources of input in the future.

VOICES is definitely a trend setter in the virtual agent space, enabling organizations to hear what their customers are really saying. It’s up to the company to decide how to use that valuable information to improve the customer experience and excel beyond the competition.

Does Web Self-Service Cause Baldness?

Quality ControlGreg Lavin of the Off Center Blog recently made a very interesting post about the downside of web self-service. In the blog he claims that poor self-service apps and shaky interactive voice recognition (IVR) systems account for about 10% of adult baldness. This is, of course, a tongue in cheek observation and not one grounded in actual research. But he almost certainly has a point.

Lavin’s comments are reminiscent of the article I discussed in my previous post on the frustration many consumers have with poorly conceived service avatars. Why do many consumers prefer talking to a real person as opposed to interacting with a virtual agent or service avatar? The answer would have to be that the human is better at understanding them and providing the assistance they need.

But web self-service applications aren’t going away. Technologies are constantly improving, and the day may come when a virtual agent can get to the right answer faster than a human. Until that day arrives, Lavin suggests that we focus on testing our service bots and other web self-service technologies. He provides some useful tips for applying quality assurance checks to existing apps.

Don’t assume virtual agents are working smoothly. Make sure that someone is reviewing logs to analyze how many consumer questions go unanswered. Encourage users to provide feedback after they interact with self-service agents. Use social channels to get input on self-service. Most of all, invest in continuously testing your agent technologies.  Testing can help you ensure that your web self-service is good for the customer and good for business.

The Evolution of Virtual Agent Technology

Evolution of Virtual Agent TechnologyVirtual agent technologies have evolved over the past two decades from simple pattern matching programs, to systems understanding natural language and intent, to extremely powerful question answering computers. All of these technologies are still available and have their place. As I noted in a previous post, local business owners can leverage simple chatbots to add spice to their websites and engage customers. More capable virtual agents that use natural language processing and search have proven their value in the customer service realm. These agents are available to interact with consumers when and where they shop, on mobile, websites, or embedded in social media channels. Cognitive computers like IBM’s Watson offer a whole new level of emerging capability. These powerful processing agents can possibly assist in healthcare scenarios, provide real time insights to financial analysts, and generally offer a new type of quick response, deeply knowledgeable consulting to aid humans in all aspects of their lives. To illustrate the evolution of virtual agent technologies, I’ve put together the Evolution of Virtual Agent Technology infographic. You can click on the above graphic to see an enlarged image.

Aivo Virtual Agents on Facebook

LuigiI recently had an opportunity to try out several of Aivo’s virtual agents on Facebook. Aivo is a company offering virtual agent technology for the customer support segment. They call their product AgentBot and the technology is available on multiple channels, including social media platforms. Two AgentBots are available to chat with from Aivo’s Facebook page.

One of the virtual agents is Luigi, a virtual spokesperson for the Fiat brand in Argentina. I like how Luigi is available as an app from the Aivo page. Just click on the app and a special page appears within Facebook with an image of Luigi, a list of frequently asked questions, and a dialogue box that invites you to ask about Fiat cars and services in Argentina. The fact that you can chat with the agent without leaving Facebook is a big plus. You can converse with Luigi in Spanish. My Spanish is pretty nonexistent, but I was impressed with Luigi’s ability to understand and respond to my questions. I asked Luigi how much a Fiat costs. I expected some sort of vague response. Instead, Luigi asked me to select from a list of Fiat models. Once I’d selected the model, I was presented with an image of the car and a base price. I also got a list of options and other features.

Another very social feature of the Luigi Agentbot app is the user’s ability to rate each of the agent’s responses with a thumbs up or thumbs down. If you choose a thumbs down, you can be more precise about the reason for the down vote. All of this information is presumably funneled back into the database to help make Luigi smarter over time.

Sofia is another AgentBot representing Telefonica. She’s available to speak with from the Telefonica Facebook page. Her chat interface is also embedded within Facebook so that the user doesn’t have to exit the Facebook landscape to engage in a conversation. I started a chat session with Sofia, but my Spanish was so lame that she quickly realized I needed special attention. She seemed to try to refer me to a live chat support agent.  She also presented me with a form that I could fill out in order to get a call back from a customer service rep.

Virtual agents that can engage with customers from within the social channels they frequent can be very compelling. These Aivo AgentBots are a good example of virtual agents designed for social spaces.

Do Service Avatars Have a Bad Reputation?

Hi I'm a Service AvatarThere are different names out there for intelligent conversational software programs that offer customer support. I typically use the term virtual agent, but the phrase “service avatar” seems to be in the vernacular as well. All Things D published an article today by Jeff Cavins, CEO of FuzeBox, in which he seems to use service avatar as a derogatory term. The thrust of the article (which certainly has a clever title!) seems to be that cloud-based services companies, in particular Software as a Service companies, have lost the fine art of true customer service.

Cavins observes that many cloud and tech companies are so focused on growing their business, that catering to the customer isn’t a priority. Being tech savvy, these companies feel that it’s okay to let the customer fend for him or herself by running them through what Cavins refers to as a “low-touch, self-service experience.” Automation has become so common in everything these tech companies do, that they’ve naturally sought ways to automate the customer support experience too. Cavins claims that customers feel abandoned and that they don’t appreciate being told to “deal with my service avatar.”

Those of us who believe in virtual agent technology might have our feelings hurt by this criticism. But I’ve long been told to take criticism as a gift. It would be great to start collecting objective evidence from consumer interactions with virtual agents to understand what is and what isn’t working for customers. Will a consumer always prefer interacting with a real person over chatting with a service avatar? Are there settings in which a person might actually prefer talking to a virtual agent? If a consumer can get to a virtual agent is seconds to have a simple question answered, wouldn’t that be preferable to waiting many minutes to speak to a human service rep? Is it possible that a person might feel more comfortable conversing with a service avatar about money matters or health issues? Perhaps there’s a threshold of technical capability that has to be reached before a person will really prefer a virtual agent over a live human? For example, if the service avatar / virtual agent is able to process more information more quickly and retrieve the correct answer more reliably than a human, wouldn’t the consumer prefer dealing with the agent?

These are all areas ripe for study. In the meantime, we should take the observations of Jeff Cavins to heart. Let’s not assume that, just because it’s easier and cheaper for us, the consumer will always be fine chatting with a bot.

What Does Cognitive Computing Mean for Virtual Agents?

Watson2Forbes recently published an article on IBM Watson and the intent to use Watson’s cognitive computing abilities to provide customer help desk services. In a previous post, I wrote about the technology behind DeepQA, which is the question answering framework used to power IBM Watson. The DeepQA technology seems like it could be a game changer for the virtual agent landscape.

How does cognitive computing push the envelope? DeepQA / IBM Watson is not constrained by the limited amounts of information that most current virtual agents have access to. This is an evolutionary process. Chatbots have been far outstripped by intelligent virtual agent technologies over the past decade. Chatbots are simple pattern matching devices that can only respond to a question that has already been programmed into their database. Virtual agents don’t have to be primed with all the questions in advance. They have the benefit of speech and/or language recognition, natural language processing, and search. They can understand a person’s intent and search a limited database or set of website content to find an answer, primarily based on keyword matching.

Cognitive computers like DeepQA go even further. Such question answering systems can be primed with almost unimaginable amounts of information. IBM Watson can be given access to every bit of internal product and company documentation, as well as online review sites, analysts articles, and on and on.  This is far more information than a traditional virtual agent, or even a human, could consume and process. Relying on its massively parallel probabilistic evidence-based architecture, IBM Watson can very quickly find possible answers to almost any question, determine which answers are most likely to be correct, and offer a response.

What do these newly emerging question answering technologies mean for the future of customer service virtual agents, personal digital assistants, and web self service as a whole? It remains to be seen how products such as an Ask Watson will perform in the real world. It’s also not clear if Ask Watson will be cost competitive when compared with more traditional solutions.

How the virtual agent landscape develops will depend not only on emerging technologies, but on what consumers expect from the systems they interact with. We’ll be keeping an eye on developments.