I thought the most interesting session was on agents and automation that discussed how speech-enabled applications can run in the background to support the agent. In this session we heard speakers from Genesys Labs, IBM Research, and Convergys, who talked about using voice to streamline workflow, such as adding a VUI to CRM GUI in order to voice navigate screen actions. Talk about a great use for speech. In one demo during the session it showed how keywords spoken by the agent, would trigger actions to happen, agent screen fields to be filled in, or different screens to appear. A simple example would be, at the end of the call the act of the agent saying “thank you for calling”, would trigger the application to finish that caller, close out the record and bring up the next call.

Similarly, another speaker talked about speech-enabled “call center agent buddies” who listen in on the call and offer up suggestions for what the agent should say, or bring up content related to what the caller is talking about, delivering information to the agent as they need it. In the demo of this, the application also created a log of what happened during the call that the agent could go back and edit later, greatly reducing the workload on the agent by eliminating a lot of the previous manual process they had gone through, and improving accuracy.

The second speech aspect of contact centers that was explored was in a session that included Nuance and Cisco, on how we can improve the experience of mobile callers calling into contact centers through the use of multimodal applications. In this session and others, the focus was on how to properly mix modalities in these applications, such as when its best to use speech, text, voice or video for the caller.

Another aspect of speech in the contact center that was discussed was how using voice search and business directories are making it easier for callers, particularly in the mobile space, to be connected with a contact center. The easy use of speech in directory applications not only has the capability of driving up call volumes, but also in getting callers used to speech, and used to “saying what they want”. A double edged sword, directory voice search will bring us new customers, but puts even more pressure on application designers to do a superlative job in designing applications that meet caller expectations as a result. In addition, it was discussed that as a result of increased call volumes there is an even greater need to use speech recognition and text-to-speech to automate more of the resulting calls.

Lastly, it was the use of speech analytics to improve contact center performance in general. Speech analytics has been around for awhile, but is starting to gain a lot of traction. Speakers gave examples of using speech analytics on recorded calls to both uncover business processes outside the contact center that impact call volume and customers satisfaction and within the contact center focusing on agent behavior or customer satisfaction. For example, one customer uncovered a formatting problem with bills which was driving customers to call in droves, but was something that the contact center had no reason codes for in the system. Without the reason codes there was no trail back to the problem, but speech analytics nailed in instantly.