By Mike Aoki
Speech analytics have progressed tremendously since their early days as little more than a way to measure vocal “tone” and single keywords. Now some programs can convert the content from customer interactions – calls, chats, emails, surveys and social media – into a format that can be analyzed. That gives you “the ability to monitor, record and transcribe 100 percent of customer contacts across different channels”, according to speech analytics expert, Steve Wilton, Principal Consultant at Wilton & Associates Inc. (http://www.wiltonassociates.ca).
Wilton notes that speech analytics have progressed from phonetic systems that recognize single keywords to advanced LVCSR (Large Vocabulary Continuous Speech Recognition) programs. They generate a transcript of the conversation and uses context to determine which are the right words and phrases. That provides more accurate results.
Wilton reports the latest developments in speech analytics include real time analysis to identify actionable insights during the calls rather than after they end. Another key trend is the use of artificial intelligence (AI) and machine learning to continuously improve data analysis. Those self-learning algorithms will improve at identifying key data trends to generate more accurate results.
DMG Consulting LLC predicts speech analytics demand will grow by 17% in 2018. Here are several examples how advanced speech analytics are helping contact centres and customer-focused organizations:
Your CEO wants to know how many customer inquiries are caused by a particular product. Can you easily find that information? What if the sales and marketing team asks about responses to their new ad campaign? Can you provide verbatim customer quotes?
There are now speech analytics programs can recognize the context of conversations. They can tell you how many customers sounded excited about a product or talked favourably about a particular ad campaign. That enables you to provide near real time information about your contact centre’s customer interactions.
You can track upsell opportunities by product type to see if agents favour certain items. Many systems analyze an agent’s vocal tone to determine if they sound confident when talking about sales. It can identify periods of silence caused by an agent hesitating to suggest an upsell or pausing before tackling an objection. It can even flag when a customer has to repeat themselves. Some systems track the most common customer objections to a sale so, you can train and coach your agents to deal more effectively with them. Those are powerful tools for improvement!
Another benefit of speech analytics is the intelligence it provides to your product development, sales, marketing and C-suite teams. They discover why customers are cancelling, what retention offers are most successful and if agents are showing the necessary empathy and customer service skills to drive higher NPS (Net Promoter Scores.) Better yet, this data is based upon ALL calls versus traditional call monitoring where only a tiny percentage of calls are analyzed and fewer still yield Voice of the Customer (VOC) data.
Coaching and QA
Speech analytics helps with coaching. One of my biggest frustrations as a coach and trainer is finding good customer interactions to review. I usually review a dozen or more calls and emails to find one good coaching example.
How does that affect your team leaders and/or quality assurance (QA) team? How many interactions do they review before finding a good coaching example? How much time does that waste and how much does that cost in lost productivity?
Imagine being able to find the right example just by clicking a few keywords into a program. How much time would that save? How would having the right example boost your coaching conversation?
The right speech analytics software can sort interactions by various parameters such as interaction type e.g., cancellation, renewal, upsell, agent behaviour (hesitation or stress in the agent’s voice) and customer behaviour (threatening to escalate to social media.) You can identify where your agents need help and quickly retrieve good examples for coaching.
Another benefit is that ALL calls can be analyzed to find overall trends. That combats the common agent complaint of, “That call was a one-off. I usually don’t sound like that!”
Wilton points out that with speech analytics a team leader or QA coach can offer real-time support while the call is still occurring. They might be able to instant message the agent with key diagnostic questions or the answer to a customer inquiry. That can result in improved customer service.
Speech analytics software can conduct automated QA scoring. Wilton notes, “organizations easily can reduce QA staff by 50% using speech analytics for this purpose. Speech analytics is a better option for agents who now have 100% of their calls scored (not just 1%-5%.). This is done in a timely fashion (often in hours rather than weeks), and it is done objectively, by a machine, so there is no opportunity for believing there is a bias.” That can improve coaching consistency and agent performance. This analysis can also highlight training opportunities and process gaps.
Speech analytics can reinforce compliance. If you are in a heavily regulated industry, speech analytics can identify when agents veer off the mandated regulatory script. It can identify interactions in which the agent failed to use proper compliance language or omitted key phrases. That allows you to coach back to those issues. It also provides you with a mechanism to gauge compliance, e.g., “Our agents follow the compliance script for new sales 99.7% of the time.” A good system can also flag individual calls when that does not occur, giving you the opportunity to address those issues in real-time.
What should you ask, if you are planning to add or enhance speech analytics in your contact centre? Wilton says, “Some speech analytics companies design their solution to be self-sufficient – with vendor support as a fraction of their revenue costs – while others charge significantly more” to get fully up and running.
He adds, “Step-by-step guides that walk (your contact centre) through how to achieve Return-on-Investment (ROI) and having expert one-on-one support” are important for a successful speech analytics implementation.
Wilton gives one final tip, “Ask if a vendor offers the ability to ‘test drive’ or do a ‘proof of concept’ for a fraction of the price compared to purchasing it. This allows a company to see what it can do while being supported by a team of contact centre and speech analytics experts.”
Mike Aoki is the President of Reflective Keynotes Inc. (www.reflectivekeynotes.com ), a Canadian training company that helps contact centres improve their sales and customer experience results. A contact centre expert, Mike serves on the Advisory Council of GTACC (the Greater Toronto Area Contact Centre Association www.gtacc.ca )
Mike Aoki is the President of Reflective Keynotes Inc. (www.reflectivekeynotes.com ), a Canadian training company that helps contact centres improve their sales and customer experience results. He serves on the GTACC (the Greater Toronto Area Contact Centre Association www.gtacc.ca )