Prediction: 10 Ways Speech Analytics Will Change The CX Landscape 🦾

What's up CEXy people 🤓 we hope you've all had a productive week. We definitely have!

This week's newsletter is all about speech analytics. Like it or not, phone calls are here to stay. But the way we use call data is beginning to change.

But before we get into it, we've got some cracking job interview advice for ya'll!

INTERVIEW ADVICE: A GUARANTEED WAY TO LAND THE JOB

• Interviewer: "Before we get started, feel free to pour yourself a glass of water"

• You: *Pour water into the cup until it begins to overflow 🫗*

• Interviewer: "Oh, are you a little nervous?"

• You: "No, I just always give things 110%" [mic drop 🎤]

10 ways that speech analytics will change the world of CX

This week, we stumbled upon a recently published interview between Steve Jobs and Joe Rogan. As legitimate as the interview sounds, it was produced entirely using artificial intelligence (AI) which blew our mind 🤯

This got us thinking about the recent developments in speech-to-text technology and subsequently allowed us to come up with 10 predictions about how speech analytics technology will change the CX & EX landscape in the next 3-5 years.

Prediction 1: The contact experience score (CXS) will soon begin to sit next to NPS on the corporate scorecard 📈

On average, the typical customer survey yields a response rate anywhere between 5-15%. What if your organisation could have a response rate of 100%?!

Organisations like Medallia currently have software which objectively scores 100% of calls, producing what is known as the Contact Experience Score (CXS). This score is derived by applying speech & text analytics to each respective call transcript, analysing factors like sentiment, emotion & themes to produce a composite score between 0-100.

As this technology advances further (and ultimately becomes cheaper to use), we predict that the CXS score will become a metric that is reported to the C-Suite for 2 reasons;

1. It is an objective way to measure the agent-customer interaction as opposed to surveys which is the subjective view of one party.

2. It is representative of 100% of interactions* that come into the call centre (versus a survey which only captures feedback on 5-15% of interactions.

*CXS can also be produced for webchat, email & social media responses.

Prediction 2: Speech analytics will automate the manual QA process ✍️

In today's world, majority of contact centres have a QA team who manually review 2-3 calls per agent a week. After listening to these calls, they fill out a scorecard which "objectively" assesses the performance of the agent according to a specified criteria e.g. script adherence, soft skills & compliance.

In the next 3-5 years, we anticipate that the manual effort associated with QA will be automated using speech analytics. There are 3 reasons why this makes sense;

1. Speech analytics will be able to review 100% of contact volume. This would be a 98% uplift on most current QA procedures.

2. Automating the scoring component of QA will not make the role of QA teams redundant. It will just allow them to focus on coaching & development; something which is more more valuable to their business.

3. QA scorecards are generally static in nature. This makes them easily programmable for speech analytics to assess & review.

Prediction 3: VoiceBots will replace contact centre agents for routine tasks 🤖

Chatbots are replacing actual agents for mundane, repetitive tasks, right?

There's no reason why VoiceBots will not replace agents for routine phone calls. This is already (kind of) happening with smart IVRs however we anticipate that there will be full-blown conversations where customers will have no interaction with a real person.

Don't believe us? Check out infinitus.ai who claim that their platform can automatically "handle volumes of tedious, time-consuming phone calls with super human accuracy and easy to track updates" 😲

Prediction 4: Customer churn will be predicted in real-time 🤠

In last week's newsletter, we briefly spoke about the emergence of "next best action" technology. In order for next best actions technology to work, it must be able to ingest and digest large volumes of data.

We predict that real time speech-to-text transcriptions will be the catalyst for widespread adoption of predictive & next best action technology.

This is how we envision it will work;

• Customer calls contact centre // speaks to an agent (or bot)

• Conversation is transcribed from speech-to-text in real-time

• Text analytics will be run on top of the transcription, identifying topics, sentiment & emotions of customer

• Outputs will be fed through a machine learning algorithm

• Algorithm will predict future behaviours according to outcomes observed in situations similar to this one

• Predictions will be fed back to the contact centre / digital teams, providing them real-time recommendations on how to resolve this customer

The above dot-points will all occur within milliseconds of each other, providing agents real-time insights and recommendations of how to proceed with the call.

Prediction 5: Sales teams will be automatically notified of upsell / cross-sell opportunities 💰

Speech analytics will provide the capability for different teams across the business to be alerted of customer needs, even if the customer has not explicitly requested something.

For example, a customer may contact the Car Insurance team and mention that they're about to purchase a new house. In the current world, the agent would need to recognise the cross-sell opportunity and alert the Home Insurance team.

In the future, speech analytics will be automatically able to identify the cross-sell opportunity, triggering an email alert to the respective Sales team within the Home Insurance department.

Prediction 6: Businesses will be able to close the loop with the most important customers, not just the loudest 📞

At the moment, the only customers who receive a phone call to "close the loop" are those who provide a low NPS score, or leave a scathing comment.

Speech analytics will change this. By analysing 100% of phone interactions, businesses will be able to close the loop with customers who are most "at risk". Speech analytics will be able to pick up on certain language or topics used within conversation which could immediately raise a red flag. These could include;

• Vulnerable customers

• Repeat callers

• Churn Risks

• High-value customers who are frustrated

The ability to proactively identify these "at risk" customers before they complain further will lead to a reduction in repeat contacts and less dissatisfied customers.

Prediction 7: Agents will be able to self-learn after each call 👩🏼‍💼

Currently, agents have limited ability to receive feedback on how they're performing. They are reliant on receiving QA reviews (which come through sporadically), or they could get visibility of survey feedback provided by customers who have had an interaction with them.

In the future, speech analytics will provide agents with insights after every call that they take. Automated Quality Management scores will allow them to understand which calls went well, and which calls require further improvement.

In addition to the above, they will be able to receive guidance on soft skill metrics such as;

• Politeness

• Helpfulness

• Empathy

• Knowledge

• Professionalism

This real-time feedback loop will keep employees engaged and will allow them to continuously iterate their approach according to the feedback that they receive.

Prediction 8: Regulatory non-adherence can be rectified immediately ❗️

In today's world, most call centres rely on manual QA processes to uncover regulatory non-adherence. For example, QA teams within the Insurance industry are responsible for flagging certain occasions where an agent does not explain the duty of disclosure correctly to customers. This approach is fraught with danger as the QA team do not have the capacity to review every call.

In the future, speech analytics will have the capability to monitor 100% of interactions and automatically alert QA or compliance teams of non-adherence. If businesses can quickly identify non-adherence before it becomes a systemic issue, this would lead to substantial cost savings by avoiding unnecessary fines.

Prediction 9: Disposition codes will actually become useful again 🧑🏻‍💻

For those who have ever worked in a contact centre, think about how frequently your disposition codes were changed. We bet that it was rarely, if ever at all.

With speech & text analytics, this will change. Disposition codes will become dynamic in nature. AI & Machine Learning capabilities will be able to identify emerging themes AND will also be able to retroactively quantify the volume of a particular topic.

For example, an organisation would be able to leverage speech & text analytics to identify volumes of calls which mentioned COVID even before a disposition code was created for COVID. How good!

Prediction 10: Customer surveys will become more targeted in their approach 🎯

With the evolution of unstructured data analytics, what does that mean for the good ol' customer survey?

We still expect the customer survey to remain a cornerstone of customer experience however it will be used differently. The questions asked will be more granular in nature, aiming to understand the "why" behind a certain customer behaviour.

Touch-point surveys i.e. those triggered after certain moments across the customer journey will also remain critical particularly across digital channels.

Graph of the week 📈

The double diamond; a design framework to ensure iteration & continuous improvement. Definitely one of our favourites!