Speech goes mainstream

Sabio’s Founding Director Adam Faulkner believes that 2006 is proving to be the year when speech technology has gone mainstream, with organisations now looking beyond the technology to establish how speech-enabled customer service processes can add value for both the customer and the organisation itself.

The first step is to make sure you know who your customers are. They aren’t all the same, and it’s essential to know who they are and why they’re calling you before you can start to select an optimum speech service strategy. If you’re developing a public service access solution for a local authority, for example, and you know that Council Tax issues account for 55 per cent of your calls, then it makes sense for Council Tax to be a key focus of any greeting. Having it sit down at number five in a lengthy list of ‘press x for y’ dialogues just isn’t going to work.

Secondly, remember that all efforts to re-engineer your contact centre to improve your customers’ experience, should always follow a clear ‘don’t make me think’ philosophy, where the service or application is effectively transparent for the caller. Lots of organisations still get this wrong – you only have to visit websites such as www.gethuman.com that give instructions for bypassing technology systems to get through to a live agent, to realise that consumers are generally frustrated with poor quality self-service and prescriptive service applications.

What are the most common misunderstandings when it comes to speech recognition?

Contact centre speech application designers should start off by getting away from the traditional auto-attendant mindset, and actually develop a customer ID approach that doesn’t result in three minutes of dead time for the customer. Next you’re faced with the challenge of how to route that ID’d call effectively – in a way that adds to the customer experience rather than simply taking them through the whole process again. A key focus here should be on finding out what callers actually hope to achieve from their call. If, for example, a user calls a bank for a balance enquiry they don’t want a long conversation, they just want secure access to their account details. Automation can have a key role to play here, identifying callers, providing a secure identification and verification process, and providing organisations with the ability to route calls based on previous customer experiences.

Often the differences between a poor and a more satisfactory customer experience aren’t that great. At the most basic level, customers don’t want too much. They want easy access to your products and services – they’re used to 24/7 web availability now, and don’t really see why they shouldn’t get the same access to your products and services through other channels.

In what ways do you see speech technology changing communication needs for customers and contact centers?

Every year for the last ten years, pundits have been busy predicting that the mainstream adoption of speech technology is just around the corner. Now, however, the predictions are proving right. From our own project-based experience, we’re finding that more and more organisations are moving beyond initial speech trials to actually incorporating speech components into some of their core customer service processes. Instead of just implementing the technology, the organisations who are getting it right are the ones who are addressing key customer service questions – such as ‘who are our customers’, ‘why are they calling us’, and ‘what’s the most appropriate way to handle this call?’

Getting the three key customer requirements – recognition, routing to the correct agent, and availability – are all essential components of any successful customer experience. For the customer getting these right might seem simple, but at any given point, such a transaction could involve a range of different systems and applications including speech recognition, integration with CRM (Customer Relationship Management), dynamic skills-based routing, effective workforce management and quality monitoring – and that’s a lot of technology to deal with.

Where do you foresee speech technology five years from now?

The majority of organisations now understand that self-service doesn’t simply equate to poor service – it’s increasingly the UK consumer’s preferred way of doing things. By 2010 this will have gone further, with customers starting to use Web-based consoles to manage their own service accounts. More and more organisations will use live agents simply to deal with service exceptions, managing more complex customer interactions and providing expert support for customers who are escalated out of the self-service infrastructure. This is going to have an inevitable impact on the make-up of the typical contact centre of 2010.

Self-service applications will be primarily speech-based with customers accessing them through their home and mobile devices. Presence technologies such as SIP will determine how and where a customer is, before presenting appropriate dialogues and functions depending whether they’re on the move or not.

For speech application designers, many of the challenges will be the same as today – making sure that customers don’t have to wait too long to have their queries answered, that customers are recognised when they call in, and that records are available during the call. These core challenges will be as equally valid in 2015 as they are now.

Do you feel that speech technology will eventually replace the contact center agent?

Every day millions of users worldwide are successfully using speech-based technology to transact with organisations. Some are calling directory enquiries, others are getting their latest account information or accessing their personal records. For many consumers, these speech-based customers experiences are extremely productive and wouldn’t be improved by the addition of a live agent.

2006 has seen a major shift in the adoption and deployment of speech technology by UK contact centres, with a growing number of organisations starting to deploy speech applications to support their core customer service processes. For many this is a fundamental change, as it involves moving beyond internal concerns such as potential cost savings or cutting the number of calls into the contact centre, and instead thinking about the suitability of a transaction for automation, and how likely customers might be to accept speech as an effective means of resolving their call.

Not all applications are immediately suitable for speech-enabling. For example, customer complaints procedures will not be suitable for automation because they typically need the adaptability of a human to resolve those complaints. Order processing, payments and information services on the other hand are more readily migrated to self-service. So while speech applications can be leveraged to achieve significant benefits for the customer and for the business, they are unlikely to take over completely.