Real-time speech analytics: Super-powering sales and service?

10th Mar 2014

Speech analytics may have only recently become commonplace, but with commercial products having been on the market for over decade , it’s perhaps little surprise that the next generation of speech analytics solutions are already emerging. And the next generation is real-time.

Real-time monitoring tools have been made possible by the technological advances of the last few years, and even though they still require significant processing power, for those first-movers willing to do the heavy lifting, they promise an array of benefits.

“It’s in the early days,” notes Art Schoeller, principal analyst at Forrester. “But we’ve been striving for this for a long period of time, and now that we have bigger, badder processors and better algorithms, it has become more feasible.”

So what’s the difference between real-time monitoring tools and the speech analytics that is most commonly implemented in contact centres?

Traditional speech analytics turns unstructured information from customer phone calls and translates it into structured information that can be searched and analysed. But all of this is done after the phone calls have been completed – it is post-call analytics.

The latest breed of speech analytics tools, however, takes place in real-time – enabling action to be taken while a call is still live.

With analytics able to be applied in real-time, a whole range of dials can be turned to improve numbers, depending on where in the contact centre the solutions have been deployed. Compliance, sales and customer service can all be positively impacted by the presence of real-time monitoring, enabling supervisors and agents to ensure a positive outcome.

Customer service

Bad customer service is costing US companies $41 billion a year, according to research by NewVoicemedia, with 58% of those polled never returning to a company that has let them down. Similar research by the US Chamber of Commerce indicates that 68% of customer defections are directly related to poor service.

In light of this, being able to identify customer language that has the hallmarks of a disgruntled client likely to churn, and being able to respond by offering retention offers, all within the same call, is something that could potentially have huge financial benefits for businesses. Real-time speech analytics promises just such a solution.

“You can monitor in real-time what the agent is doing and what the customer is saying, and you can use that to drive actions,” says Jim Davies, research director at Gartner. “So if the agent is saying the wrong thing, or the customer seems to be irate, then you can either alert a supervisor who can start to monitor the call or jump into the conversation, or you can start screen popping the agent with some information.”

Omer Minkara, senior research analyst in the customer management technology practice at the Aberdeen Group, adds: “It is valuable in terms of customer sentiment. If the agent is speaking with a customer and the voices are overlapping, the chances are it is indicative that the two are not listening to each other, so the agent can be prompted to stop and listen to the customer. Or if the customer uses particular words that are indicative of frustration, then the agent can be prompted with alerts so that they know how to respond.

"Without those real-time capabilities, you can still run post-call analytics and measure agents’ performance, but you can never get back to the moment of truth – that is, the second that you were engaging that customer and how you could have made it up for the customer in an effective way.”


Real-time monitoring also has implications for sales, effectively ensuring that upsell opportunities are not missed by agents. And when used in combination with wider customer data and agent notifications, real-time speech analytics enables agents to deliver customised sales offers that are more likely to resonate with the customer.

Davies provides a good example of this procedure in action. “say a customer is on the phone talking to their bank, asking about their bank balance, and while the agent is looking it up the baby starts crying in the background. The agent asks is that a baby he can hear and the customer says ‘yes, I became a dad last week’. As soon as he has said that, the agent could get a notification on his screen prompting him to offer the customer life insurance.

“This is not something that the agent may have thought of, but because the customer is talking about being a father, and having a baby, and parenthood, those key words tell the analytics system that the customer is much more likely to take out life insurance because he has to take responsibility as he’s a parent. So that screen pops the process flow to the agent that they wouldn’t necessarily have picked up on if left to their own devices.”

Furthermore, real-time speech analytics can enable businesses to improve call efficiency and efficacy, by providing agents with guidance on disclaimers and close procedures, thereby minimising the number of disputes and returns.


Arguably the biggest driver for real-time speech analytics is in the area of compliance. Violations of government regulations can be costly for businesses, so real-time monitoring can help to ensure that they keep to the guidelines via timely script reminders. And in the event that a violation that occur, the agent and the supervisor can be notified immediately so that they can take rapid steps to reduce the likelihood of a complaint.

“The biggest use case at the moment is in the compliance area,” suggests Davies. “It has a big role there because if you can address non-compliance in the same call it makes life so much easier. If the agent has to follow certain steps to be compliant, and they miss out one of them, the speech analytics system can see it because certain words and phrases weren’t spoken. So the system can screen pop the agent to say don’t forget to do a particular thing.”

“The money is in the fraud and risk and compliance area, so that’s really providing the engine to fund deployments of real-time speech analytics,” agrees Schoeller.

Real-time revolution?

While still a relatively new tool, real-time speech analytics solutions are now being offered by a growing number of vendors, including the likes of NICE, Verint, CallMiner, Interactive Intelligence and Castel. And even though the proportion of speech analytics implementations that have real-time functionality is still fairly small, large enterprises such as Santander Consumer USA are already buying into the technology. For speech analytics, the future is real-time.

“Only approximately 10% of speech analytics users have real-time functionality according to our research," notes Minkara. "That’s because it requires a lot of heavy lifting in the back-end; substantial data processing power is needed in to analyse speech data on a real-time basis. However, real-time speech analytics empowers agents with distinct advantages.”

Replies (0)

Please login or register to join the discussion.

There are currently no replies, be the first to post a reply.