This article originally appeared in the Oct. 2011 issue of Customer Interaction Solutions
Autonomy (News - Alert) helps companies understand their vast and growing amounts of human-friendly or unstructured data such as text, e-mail, web pages, voice, or video. The company got its start offering technologies enabling enterprise search, helping knowledge workers locate unstructured information inside enterprise repositories like intranets, document management systems and even file shares.
However, Autonomy grew to power other applications, which now include protecting organizations' information assets with e-discovery, content and records management, compliance and legal technologies. The company also provides solutions to both facilitate and analyze customer interactions from the web, contact centers and social media. Autonomy also OEMs its technologies to more than 400 software companies, so it's likely your business is already using Autonomy-powered solutions. CIS magazine recently spoke with Andrew Joiner, CEO Promote, Multichannel Technology at Autonomy, to find out more.
Who are your target customers?
Joiner: Globally, Autonomy has 25,000 customers across every industry vertical. In the contact center space, specifically, we work with the largest and most sophisticated centers throughout the world.
What is meant by speech analytics?
Joiner: Speech analytics is a term used in the contact center space to mean the analysis of recorded conversations between customers and employees, though speech analytics doesn't have to be defined that narrowly. What's interesting is that conversations tend to be long and sometimes winding, because, in contrast to computers, humans communicate in an inexact way. But, these routine conversations have extraordinary value. Customers tell us all kinds of things that are profoundly useful as we manage our businesses. Sentiment, preferences, competitive intelligence, multichannel challenges and future plans are just some of the nuggets our customers find through speech analytics – nuggets, incidentally, that divisions outside of the contact center are starving for.
How does it work?
Joiner: Autonomy has an engine that can process and understand all unstructured information, which includes call recordings. It has over 500 functions that help illustrate the concepts expressed in a phone conversation.
Using advanced techniques like clustering, sentiment detection and automatic classification it can easily surface hot and breaking issues throughout the contact center. It can determine when customers are upset or frustrated and why they’re calling. The engine looks at the words in conversations contextually, so the results are far more relevant than other approaches, helping automate the process of root cause identification and the recognition of emerging patterns.
How do customers typically use Autonomy Explore?
Joiner: Autonomy customers don’t have to change their processes or underlying call recording technologies. Autonomy provides a layer that sits on top of the infrastructure while reading and listening to the information, the way a human would. Once it forms an understanding, marketing, customer service, product management and other groups can tap into it. The value explodes from there.
For example, a consumer may call about a defective product. Autonomy Explore analyzes what is being said and categorizes the conversation and cross-references the concepts with other recordings, web visits, e-mail correspondence and even social media buzz. From this information, Autonomy Explore searches for patterns like changing call volumes or changing sentiment for a competing product. It automatically generates alerts and reports in real time to flag the situation and address it before it worsens.
Provide an example of how one of your customers is benefitting from speech analytics.
Joiner: Our best customer examples tend to come from customers that subscribe to our vision of multi-channel interaction analysis – those companies that apply the solution broadly. One of our biggest customers used our platform to identify cell phone fraud because the system was able to flag suspicious calls where customers where providing the same pattern of phone replacement requests. This trend was not statistically significant to the volume of calls, but it was an unusual pattern and required the company to ship new phones, which is expensive. Ultimately, they were able to involve the authorities.The customer estimates savings in the tens of millions.
What steps should businesses take to ensure a successful speech analytics implementation?
Joiner: Firstly, I’d recommend that they select a conceptual-based technology that does not rely on keywords or transcriptions. This will simplify the process of building topics and automate the discovery process as related concepts will be generated by the system, while ensuring a focus on relevance rather than accuracy. When dealing with unstructured data, accuracy is not the target.
And secondly, don’t stop with analyzing voice recording. Customers interact with a brand across multiple channels to research, purchase and service their products. Focus on this challenge through a multichannel prism to ensure a complete and unbiased view of the customer, throughout the buying cycle. And, of course, give Autonomy a call!
Edited by Stefania Viscusi