IBM Watson Engagement Advisor for Smarter Customer Service

Recently my colleague Mark Smith wrote about the IBM Watson platform. Mark is our expert on technically complex subjects like IBM Watson and‎ cognitive computing and the value it can provide to organizations and wrote an educational white paper on the topic. In fact IBM Watson was awarded the 2012 Ventana Research Technology Innovation Award. I focus on the customer and the customer experience, but I became engaged with the launch of the IBM Watson Engagement Advisor, which uncannily brings the two together.

I conclude from my research into customer experience management and discussions around the customer experience that consumers want three related things from companies: to recognize them as individuals, to handle any interaction within the context of their overall relationship and previous interactions, and to provide answers that are personalized to their individual needs. Furthermore, as I described in my recent blog about the 2.0 customer, they are increasingly likely to interact with companies through smart mobile devices. Added up these requirements present a major challenge for companies, and the solution lies in data. Organizations have ever increasing volumes of customer data, found in records in CRM, ERP and other business applications, letters, forms, email, call recordings, scripts collected during Web and chat sessions, text messages, video recordings and now social media posts. These all add up to what we call big data, and it comes in structured, semistructured and unstructured forms. To provide a personalized, in-context response to any interaction, companies have to make sense of all this data and build applications that follow typical customer interactions (for example, a request for information, a billing inquiry, a sales inquiry or a complaint) and use it as the interaction is taking place.

This is the context in which IBM Watson Engagement Advisor debuts. It uses the Watson platform to analyze all available data, make sense of it and provide information back to the application the consumer is using to interact with the organization. Its built-in natural-language processing engine allows it to extract relevant words and phrases and combinations of both from text-based input and to search for relevant data in customer records, documents and other relevant sources. To use IBM’s favorite word, Watson is smart, so once it has followed a process, it learns from that and can improve how it carries out the process in the future. The consumer can thus interact using natural language and receive answers that are personalized and in context, and which should improve in both senses over time and experience.

I recognize that companies may have difficulty understanding this new approach to customer experience management. To engage them IBM has announced an early customer adoption program in which it will provide support and guidance on what interactions are appropriate to automate in this way and how to configure the technology and gain access to the data sources; the target of the program is to have an initial system running in six weeks. My recommendation here is the same as I made last year when several vendors announced tools that help companies build mobile customer service apps: Think of the customer first and build apps to match what they want, not what you think will improve the efficiency of your interaction-handling.

It is often said that customer service is the only true differentiator in competitive markets. The challenge I see is that the boundaries between marketing, sales and customer service are blurring, and what consumers want is answers, and they want them immediately during every interaction. As data volumes and types grow, finding answers is like looking for the proverbial needle in a haystack. Watson may be a platform capable of matching up to this challenge. I recommend that companies looking to improve the customer experience take a close look at the IBM Watson Engagement Advisor.

Regards,

Richard J. Snow

VP & Research Director

SoCoCare Enables Social Customer Service

Much is being written about the impact of social media on customer service, although my research into the agent desktop shows it hasn’t reached the fever pitch that many commentators would have us believe. It is true that the number of consumers using social media and as a consequence the volume of posts are astronomical. But I wonder how many of these posts actually have to do with customer service and how organizations filter out the relevant ones to help them decide on customer service policies and the appropriate action to take.

During a recent briefing representatives from SoCoCare told me about its product Social CIM, which supports a socially oriented performance management process whose steps the company terms “listen, filter, tag, prioritize, evaluate, act, respond.” The product has all been developed from the ground up by SoCoCare and runs in the cloud. Social CIM can capture social media posts from a variety of sites such as Twitter, Facebook and LinkedIn. It then uses an in-house developed natural-language processing engine and a rules-based engine to determine the content of the post and its relevance to the user organization. In this way it can filter out what SoCoCare calls spam, which in this context means posts that are not relevant or not actionable to the organization. The product then uses the rules engine to tag the post (for example, as a complaint, a threat to leave or a positive comment) and help decide where to route it for action, typically to specialist agents in the contact center. SoCoCare says the rules engine is self-learning so over time it will recognize patterns, types of posts and actions and use these to improve the rules going forward.

The product can be set up to push posts to a specific agent or agents can use a Facebook-like user interface, which I think will help gain user acceptance of the product, to see what posts need to be handled and choose which one to process. Social CIM contains tools that support agents responding to a post. An on-screen window provides a view  of previous interactions in time sequenceso the agent can put the current post in the context of previous interactions, the rules engine can suggest a best action or prepopulate responses, and Facebook-style collaboration enables agents to collaborate on a response. These capabilities help agents respond to interactions effectively. There also are various personalized dashboards that show individual agents how they are performing against targets.

Our benchmark research into business technology innovation shows that social media is not yet as high a priority as other emerging technologies, in fact ranking last among six choices, as the chart shows. Nevertheless the use of social media is so widespread that companies need to make sense of how it fits in their overall customer service strategy. Doing this wisely requires first understanding what is happening and then taking action. With respect to social media this means first using tools that show which posts are relevant and then deciding how, or whether, to respond. In this way companies can determine the place of social media in their customer engagement processes. As they look at taking such steps, I recommend they assess how SoCoCare can support such initiatives.

Regards,

Richard J. Snow

VP & Research Director