Speech Analytics for Call Centers: The Only Limit Is Your Imagination| Eleveo

Using advanced speech analytics tools to improve call center performance 

It isn’t new technology, but many commercial contact centers are just starting to consider Automated Speech Analytics’ benefits.

The challenges facing commercial contact centers aren’t substantively different from those of their enterprise counterparts; the only difference is scale.

Regardless of whether the volume of customer/agent interactions numbers in the tens, hundreds or thousands per day, contact center success is a function of the control managers have over the quality of those interactions. Proficiency is still measured in AHT, First-Call Resolution, Compliance and CSAT, and managers still need to balance agent resources against demand.

Why speech analytics can be helpful for every contact center?

Advanced Speech Analytics is the one tool contact centers of any size can use to influence each of those proficiency metrics and virtually every facet of the Quality Management (QM) process.

It’s generally accepted as best-practice technology, but small contact centers are a little late to the game. The question is, why?

A small operation may lure a manager into believing that, with direct feedback from his agents or a review of a small sample of calls, he has his fingers on the pulse of QM. Frankly, that’s not true. He can’t manually collect and analyze a large enough sample of agent/customer interactions to gain a thorough, nuanced understanding of contact center performance, and relying on agent feedback and anecdotes is no way to reliably recognize patterns within the noise and proactively resolve emerging issues.

The fact is, any manual QM effort is as labor-intensive as it is deficient.

Advanced Speech Analytics solutions vary in complexity and the resources required to use them effectively — not every solution is suitable for commercial contact centers. And no Advanced Speech Analytics solution will do all the work for you; you’ll still need to do some listening and manually explore complementary data. That said, some solutions will do much more than others in collecting and automating data aggregation.

It’s also true that supervisors and managers are already stretched and may have little time to devote to speech analytics. The key is establishing an analytics culture within the organization, where there’s widespread visibility into the data through dashboards, where everyone is accountable, and everyone contributes to taking action to resolve root-cause issues and drive change.

What is speech analytics for call centers?

While the assortment of speech analytics benefits can’t replace your contact center’s existing QM processes, advanced speech analytics is a powerful augmentation tool, helping you mine unprecedented intelligence from every agent/customer interaction.

It analyzes voice recordings (post-call speech analytics) or live customer interactions (real-time speech analytics) by identifying certain words and phrases, either defined by the user or included in a library developed by the provider - sometimes both. It also detects emotions in the customer’s and agent’s voices during their interaction.

Benefits of speech analytics tools for call center

Advanced Speech Analytics gives managers the ability to objectively score 100% of all interactions, rather than the 1–3% typical of most contact centers. Comprehensive scoring eliminates “red herring” issues and ensures that coaching and development efforts can be geared toward the issues and challenges each agent faces most consistently.

Advanced Speech Analytics provides data for answering critical questions, like:

• How are your agents engaging with your customers?
• What kinds of words and phrases are they using?
• Are they advocating for your customers or using “powerless to help” phrases?
• Are customers using a high percentage of dissatisfaction language?
• Are your agents talking over the customer?

Reviewing all contacts with a high ratio of silence to speech, for instance, may indicate areas where agents are struggling with navigating the knowledge base, where more training is needed or where ‘cheat sheets’ might be helpful. The same capabilities can identify your consistent, top performers.

How does advanced speech analytics work?

Advanced Speech Analytics is a transformational technology for contact centers. Without getting into the technical weeds, it’s enough to understand that the tool uses standard data process algorithms to transform the nuances of speech and silence into a reservoir of intelligence about customer attitudes, agent aptitude, process effectiveness and trending issues. It does the same for chat and email interactions, still detecting attitudes and emotions through keywords and phrases.

For call interactions, analytical tools explore the sounds themselves, detecting tone of voice, speed of speech and key phrase usage. It even explores the silences, and silence can be expensive; time is money.

One scenario:

• Let’s say you have 50 agents, and their average talk time per call is 300 seconds (5 minutes).
• Through advanced speech analytics, you discover that, on average, 30 seconds per call is spent in silence.
• You trace the silence to agents being unable to find the answer to a specific question promptly, and they’re waiting for someone in a chat room to help them with the answer.
• Based on that insight from the tool, you improve your knowledge base and ensure your chat room is adequately staffed.
• As a result, the 30 seconds silence is reduced to 20 seconds. ATT is 10 seconds (3.33%) shorter and therefore cheaper. It adds up.

How to maximize advanced speech analytics benefits

The key to Advanced Speech Analytics— the essential step to extracting maximum value from it — is defining the behaviors you know impact the KPIs your business cares about the most. Once you’ve defined those behaviors, you need to translate them into pre-defined phrase categories, a so-called ‘Phrase Library.’

The Phrase Library is the raw material of Advanced Speech Analytics. If, for instance, you expect your agent to use a specific greeting within the first 30 seconds of a contact (i.e., ‘who do I have the pleasure of speaking with’), your Phrase Library should include that greeting. The Advanced Speech Analytics solution can then key on the phrase in every interaction to see if and when it was used. Another example: in this time of Covid, you may choose to key on words, like ‘lockdown’ or a phrase, like ‘working from home.’

Organizations early in their Advanced Speech Analytics journey may be unsure about which phrases to include in their library. In that case, their Advanced Speech Analytics provider should offer a Phrase Library organized into appropriate themes (i.e., General Customer Service Phrases, Payment Related Phrases) and categories relevant to a specific industry.