Are Call Centres Ready for Real-Time Speech Analytics?

Author: Simon Black | Date: 14/02/2023

Over the last decade, speech analytics has grown from relative obscurity to being the staple of any call centre technology suite. The speech analytics market is expected to grow by a staggering CAGR of 22.13% over the next five years, and for good reason: it has transformed the way in which call centres operate, and allowed managers to onboard, train, and support their staff more effectively.1

Now, call centres are anxiously awaiting the appearance of real-time speech analytics. Unlike traditional post-interaction analytics, real-time speech analytics would be able to present data-driven insights directly to agents and their managers while an interaction is still in progress. 

When combined with advanced scripting workflows, real-time speech analytics insights form a powerful tool for assisting agents and improving customer experiences.

In this article, we’ll explore what real-time speech analytics is, why it’s such a big industry talking point, and what we can do to help it become mainstream. 

Suggested reading: To learn more about how speech analytics enhances customer service delivery, read our eBook, ‘Your Contact Centre Could Operate Better… A Lot Better’.

Why should speech analytics be real time?

Speech analytics’ success is easy to see. Countless studies have reported its benefits, from reducing call handling times by up to 40% to boosting service-to-sales calls conversion rates by nearly 50%.2 Its primary benefits are:

  • Tracking performance: Managers are able to easily track their agents’ performance and offer coaching as necessary. Likewise, agents are able to regularly see how they are performing, and where they might need to improve.
  • Reducing costs: With speech analytics tools in place, managers can more effectively monitor and improve their agents’ performance. This can reduce the costs associated with agent training and churn, while improving agent experiences.
  • Reducing time-consuming QA checks: speech analytics tools reduce the amount of time managers need to spend performing manual quality assurance checks, making their working processes more efficient.
  • Reinforcing customer retention: With data-backed improvements in place, call centres will begin to see improvements in their FCRs, customer satisfaction, and customer retention rates.

So with all of these benefits in mind, why should speech analytics change? The answer is that, quite simply, it has to. Call centres are facing more pressure than ever before, with many suffering high agent churn, unmanageable workloads, and low profits. Although speech analytics can reduce these burdens, it cannot effectively eliminate them in the way in which real-time speech analytics might.

With real-time analytics, call data can be offered to agents immediately. This data allows real-time guidance tools to seamlessly assist agents through their calls, and help agents improve without the need for regular training.

This solution is known as ‘real-time agent assist’, and is generally considered to be the future of call centre technology. With real-time assist, managers are likely to see a reduction in poor mental health and stress among call centre employees, as agents are able to manage and resolve high-emotion calls more easily.

Further benefits of real-time speech analytics, when combined with advanced scripting tools, include:

  • Driving satisfaction: By using AI-driven sentiment analysis, real-time agent assist can recognise when customers are feeling angry or upset and guide agents to respond calmly and carefully.
  • Increasing profits: real time tools can recognise opportunities for upselling, and guide agents accordingly. When accumulated over time, these small opportunities can equal big profits.
  • Reducing churn: With better opportunities for development and clearer day-to-day guidance, agents will be inclined to stay at a call centre for longer. As they grow in confidence, agents can use real-time analytics less.
  • Improving efficiency: Calls can be dealt with more quickly, as agents have real-time insights about how to improve their customer service. 
  • Reducing call downtime: the easy to use single interface means there is less call downtime, as agents can smoothly move from one call to the next; all the information they need is in one place.
  • Ensuring compliance: Real-time tools can not only recognise when contact centre compliance is at risk, but work to prevent regulations from being breached.

What does real-time speech analytics look like today?

Real-time speech analytics is yet to become widely adopted. With current post-interaction analytics solutions, call centre managers must wait until the end of a day or week to upload and assess their call data. This means that data insights are applied retrospectively, rather than in the moment.

To deliver the most appropriate next step to agents during customer interactions, real-time scripting and guidance tools trigger from pre-approved processes. We call this process-driven real time guidance — and the reality is that this solution is sufficient for a majority of businesses. When used in combination with a strong post-interaction analytics offering, process-driven real-time agent guidance is enough to significantly improve contact centre performance. 

However, process-driven suggestions are not led by customer-specific analytics, so the number of possible responses will be limited. Equally, contact centre leaders will still need to manually update scripts and processes over time, as opposed to the system learning for itself. 

Real-time speech analytics promises a world in which this manual intervention is reduced, but there are still technical barriers that need to be overcome before it can become mainstream. These revolve around real time speech analytics’ accuracy, cost, availability, and functionality:


Audio disruption remains one of the biggest impediments to transcription accuracy, which typically stands in the mid-high 90s. This means that some 1-5% of an interaction data is inaccurate, and AI tools are more likely to misinterpret interaction meanings. Even data that is transcribed accurately will suffer speech analysis delays, meaning that real-time speech analytics will be not as ‘real-time’ as it promises.


Real-time speech analytics is a novel and developing field. As a result, many offerings are expensive, and it’s unrealistic for call centres to implement them. Although the cost of post-call analytics varies, it has become much more affordable over the course of the last decade; looking ahead, we can expect real-time analytics to follow the same trajectory.


Because call centres don’t make their interaction data available in real time, call data can only be reviewed in batches. This prevents AI tools from reviewing data immediately, and slows down the process of speech and sentiment analysis.


Real-time analytics are either implemented throughout a call centre, or not at all. This can be frustrating for experienced agents who will not necessarily need guidance through a call. This ‘either on or off’ system also applies to specific calls: agents may not have control over when to turn analytics on or off, which could trigger compliance breaches if customer analytics are collected without their consent.

While real-time analytics and guidance still has a long way to go, there are some strong post-interaction speech analytics solutions already on the market. For real-time analytics to reach its full potential, call centre technology developers must find innovative ways to evade these central challenges.

The improvements in your contact centres first call resolution and customer churn would be unrecognisable if agents were to be advised on the next best action.

Find out why Awaken users are increasing customer satisfaction, and in-turn increasing operational efficiency by over 80%.

Get started with your free demo today.

What barriers are real-time speech analytics facing?

It’s worth recognising that the challenges facing real-time speech analytics are not purely functional. Both call centres and technology providers have a number of hurdles yet to overcome before they will be ready for real-time agent assist.

The barriers software providers are facing

Many scripting providers don’t offer speech analytics solutions, while many speech analytics providers don’t have scripting solutions. This means call centres will be unable to extract the full value of speech and customer sentiment analytics, and won’t be able to integrate agent assist tools. To overcome this, providers should:

  • Widen the scope of what technologies they offer
  • Undertake efforts to make interaction data available faster
  • Utilise third-party technologies to overcome functional difficulties

Awaken is currently partnering with a noise cancellation software company to dramatically improve call qualities, and allow AI to assess calls more accurately. This is just one example of steps technology providers can take to facilitate the growth of real-time speech analytics technology in call centres.

The barriers call centres are facing

Call centres are also facing barriers around real-time speech analytics. These start with a real nail biter: funding. 

Unfortunately, many call centres rely on reactive rather than proactive funding. This means that, without a sufficient business case, team leaders may not have enough justification to pay for a new technology.

With the call centre market size having shrunk from £2939 million to £2582 million since 2016, it’s understandable that call centres are reluctant to start investing in the unknown.3 However, highly limited cost models are preventing revolutionary technologies like real-time speech analytics from making progress.

Suggested reading: Want to learn more about how to build an intelligent contact centre? Read our blog, ‘Building an Intelligent Contact Centre: Listen, Understand and Guide’, to find out.

How can we make real-time speech analytics possible in 2023?

Having established what the barriers to real-time speech analytics are, we can now consider what needs to be done to overcome them. The power is largely — though not entirely — in the hands of call centre technology developers to create tools that can work effectively though use of AI. This is no easy task, but is feasibly achievable within the next five years.

Call centres can also take steps to facilitate the real time revolution. This starts with investing in their IT infrastructure, monitoring their bottom line, and assessing which business cases might justify the use of additional tools. These could include:

  • Low-skill agents who are required to work on highly complex topics.
  • Working with a company that has specific and rigorous compliance guidelines.
  • New agents begin taking calls and require extra guidance.

With these business cases in hand, managers will have greater justification for proposing and investing in new speech analytics tools.

Awaken has made delivering impactful real-time speech analytics solutions its priority. With a foundation in speech and text analytics, translation, scripting, and dynamic agent guidance, we are constantly looking to optimise call centre operations through innovative technology. Rather than working against call centres, we want to work alongside them to develop and deploy the future of customer experience.

So, what next?

The big question. As the momentum towards real-time analytics increases, call centres must be the ones preparing for its arrival. Those that have improved and developed their IT infrastructure ahead of time will be in the best position to take advantage of this new technology, grow their client base, and enhance customer service outcomes.

Until real-time speech analytics becomes more accurate, affordable and available, companies will best benefit from employing call centre technologies already on the market, such as post-call speech analytics and call centre scripting tools. These help call centres to:

  • Efficiently review agent performance
  • Increase agent and customer satisfaction
  • Reduce churn rates
  • Drive profits through greater customer satisfaction

The results of using these tools may differ slightly than with using real-time assist, but they will still be impactful. Furthermore, not all use cases will need real-time analytics; experienced agents with consistent FCRs and customer satisfaction, for example, may be content continuing as they are.

In a climate of economic and political instability, it’s important to remain cautious about what you invest in. Real-time technology still has a long way to go, and, realistically, will take time to hit the mainstream. Which technology you choose will ultimately depend on your resources, agents, clients, and historic data insights.

Awaken: driving innovation in 2023

Awaken offers a range of agent guidance, scripting, translation, and quality assurance tools. We support call centres around the world by optimising their processes and driving innovation across the industry.

We have built our products around finding productive, creative solutions to each of the challenges that call centres face. Contact us today to discover how we are leading the drive towards real-time speech analytics, or to learn more about our existing services.

1  Speech Analytics Market Trends (2022 – 27) | Industry Analysis

2  How advanced analytics can help contact centers put the customer first 

3  Call Centres in the UK – Market Size 2011–2029