Contact centers across Latin America are looking to transform their customer service solutions by implementing new channels, such as chat, social media, email, video, or updated voice services, which are still hugely relevant in the region. Within that scope, speech recognition technology is often seen as a means to cut costs, improve customer satisfaction, and increase productivity in the call center. But are these solutions worth it?
Speech recognition is performed by a machine or program that can identify words or phrases in spoken language and convert them into a machine-readable format. In layman’s terms, the software detects what you’re saying and automatically transcribes your words into text. Beyond interactive voice response (IVR) systems, some of the more popular call center speech recognition applications are call routing, speech-to-text, voice dialing, and voice search.
Some companies are experiencing improved service quality due to a combination of speech recognition and IVR systems, since the automated system is available when human agents are not. More outbound contact centers are using IVR/speech recognition systems to improve their effectiveness, especially to serve clients in the collections and sales segments.
Agent Monitoring
Originating from Spain, Verbio is a speech technology provider serving contact centers in Ecuador, Brazil, Colombia, and Mexico. Verbio’s core speech recognition engine uses a database of words and phrases to create transcripts of calls. This text is then analyzed to determine what the customer has been saying or how the agent has dealt with certain situations. Therefore, the software can measure whether the agents are off-script, adhering to regulations, or not performing at a high enough standard.
“When trying to track first contact resolution, companies can analyze certain phrases like “this is the second time I called to get this fixed” or something similar,” said Octavio Maupome, Country Manager for Mexico, Cenam, and the Caribbean at Verbio Technologies. “This business intelligence can vastly improve KPIs and customer experiences, to a point that human analysis of this kind of data becomes almost obsolete.”
Another provider in the region, Nuance, uses artificial intelligence in its Natural Language Understanding solutions in order to extract meaning from a customer request. It is also used in some of the company’s Biometrics solutions in order to learn and generalize patterns based on a set of observations.
Speech Analytics
Perhaps the most intriguing speech solutions for Latin American contact centers right now are speech analytics, as they help to understand the trends in what their customers are asking, increasing the chances of success in their campaigns. “Through business intelligence, contact centers can understand things like when the best time to sell something is, when collections are easier, or which agents are performing best,” said Maupome. “At the moment, many contact centers in the region are hiring a lot of people to listen to 3-4% of all calls, which is very costly. With speech analytics solutions, 100% of the calls can be reported on, without the need for human interaction. This is why some contact centers and BPO firms are showing more interest.”
Even though the quality assurance and KPI benefits might seem obvious, it can be difficult to implement such solutions. You need the right level of processing power from your IT infrastructure, specific professional services, and each region requires certain licenses to operate them. Many contact centers are of a mind that “if it’s not broken, don’t fix it”, so are not taking these solutions on as rapidly as the providers might hope.
Regional Challenges
“Regional accents as well as non-native speakers are the reality of the environments in which our solutions are deployed world-wide,” said Marco Petroni, Senior Director, Enterprise Professional Services at Nuance. “Over the course of 20 years of deployments, Nuance has ensured that our underlying speech recognition technologies are robust to a variety of conditions, not only accents, not just for the U.S. and Latin America, but for all the languages that are supported.”
For Verbio, the process of detecting accents in Latin America has been easier due to the company’s Spanish origin, which has helped develop a base of vocal samples in European Spanish. The company uses that base to produce specific products for each market, due to the level of variation in dialect, accent, and vocabulary. “Even in Mexico, the people from the north of the country speak differently to the central states, so we have to bring in individuals from all locations to record and add their unique dialects and accents to the mix,” said Maupome.
The complications go even deeper when you consider people of different age groups. People in the younger market—while ultimately disinterested in utilizing voice for customer service—have a completely different way of speaking than the more mature customers.
Business Value
According to Petroni, the main challenge in developing self-service voice solutions for Latin American contact centers is being able to articulate to the clients the business value of a potential solution. “Generally speaking, organizations or contact centers have either little or unreliable data regarding call volumes and call dispositions,” he said. “As a result, while there could be a very compelling case to be made regarding the business value and ROI of a potential solution, one is generally hard pressed to translate proof points from other deployed solutions world-wide to the potential benefits for a specific organization or contact center.”
Ultimately, speech recognition solutions are best-suited to large-scale contact centers that want to reduce their spending on human resources, particularly in call analysis. The costs of an agent transaction versus an automatic transaction are certainly higher, but companies need to consider the cost of maintaining sufficient IT infrastructure and service fees before taking the plunge into full voice automation.
Add comment