In the contact center, reports show that 65% of agents say their cases are more complex than a year ago, and that volumes are set to increase in the coming year. These increases, along with a shortage of agents in the contact center, make it more pressing to be multilingual-ready. And while some languages are widely spoken, not every language has a large enough number of native speakers to cost-effectively serve in your contact center. Or you might encounter languages for speakers who are in high-cost labor markets.

This article was written by Jarrod Davis, Content Marketing Manager at Cognigy, a Genesys AppFoundry Partner.

In the contact center, reports show that 65% of agents say their cases are more complex than a year ago, and that volumes are set to increase in the coming year. These increases, along with a shortage of agents in the contact center, make it more pressing to be multilingual-ready. And while some languages are widely spoken, not every language has a large enough number of native speakers to cost-effectively serve in your contact center. Or you might encounter languages for speakers who are in high-cost labor markets.

In both cases, artificial intelligence (AI) isn’t a nice-to-have; it’s the only viable solution to language-related CX challenges. Let’s explore how AI can assist your human agents to help them overcome language obstacles in customer service, with a focus on the following:

  • Languages with few native speakers (e.g., Danish)
  • Larger languages whose native speakers overwhelmingly live in high-cost labor markets (e.g., German)
  • Languages you cannot offshore (e.g., Hungarian)

The challenges related to these include:

  • High cost to serve
  • Lack of qualified staff
  • Languages required in time zones outside of your business hours
  1. Automation

Contact centers have a lot of processes, most of which can either be partially or fully automated. When serving languages with a limited number of staff who can speak the language, it’s best to look at the process end-to-end. The entire process involves a certain degree of language use.

With a finite number of native speakers, using them for areas like returns, identification and verification (ID&V), account information changes, simple reservations, and more are likely wasteful. This is where conversational AI can be used to either partially or fully automate your high-volume, low-complexity issues.

Even if you only automate the ID&V portion, you’re still saving anywhere from 30 to 90 seconds on every call in that language. This allows you to free up a massive amount of agent time, while simultaneously reducing agent burden and average handling time.

  1. Real-Time Translation

Native speakers in contact centers are a limited resource and yet, every agent has the knowledge and ability to help customers. So why does language get in the way?

With real-time translation for both voice and chat, agents who don’t speak the customer’s language, or who only partially do, can instantly serve customers and put their knowledge and skills to use while assisting customers quickly.

When a customer initiates contact via either voice, messaging or chat, the AI agent automatically identifies their language and intent. For text-based communication, the AI agent provides bidirectional translation, for example between Turkish and English. This is also available for voice for times when, for example, a Turkish-speaking customer calls on the phone and there is no native speaker available to help.

Using Speech-to-Text (STT), the customer’s speech will be turned into text, translated into English and passed to the agent who’s working out of a normal chat window — and using agent assistance. When the human agent responds, their answer is then translated into Turkish and finally back into voice via Text-to-Speech (TTS). Large language models (LLMs) can be incorporated, if desired, to rephrase and formulate sentences or aid in translation.

Another option is to insert background “contact center noises” like typing into the audio stream, which both masks latency from third-party services and makes the overall interaction more natural.

Finally, major advances in TTS that offer near-perfect human voices combined with LLMs, translation and AI, mean any agent can help any customer — independent of language and channel.

  1. Agent-Assist Technologies for Non-Native Speakers

Often, contact centers have agents or new hires who speak several languages, but not at a native level. They might be able to engage with a customer on certain issues but don’t have the language skills to converse about more complex issues. This also can be difficult for agents who are dealing with back-to-back interactions for hours on end.

By using AI-powered agent-assist technologies, organizations have a range of options to take full advantage of human agents’ skills, raising their confidence and reducing errors.

  • Show a translation in parallel on the agent desktop
  • Blend of conversational AI and LLMs provides suggested responses in the foreign language, boosting the agent’s confidence and fluency
  • Live transcription ensures the agent still has access to sentiment analysis, for example
  • Automated call wrap-up with summarization and case creation the agent’s working language so no manual translation is required

AI and the Multilingual Contact Center

Leveraging AI to overcome language barriers is a necessity in the contact center. The unique combination of conversational AI, generative AI and translation technology allows you to overcome the challenges that come with multilingual service. These AI solutions not only mitigate the high costs often associated with multilingual support, but also empower contact centers to operate more efficiently and inclusively.

Investing in AI to overcome language obstacles isn’t just a strategic move — it’s becoming an essential element of maintaining competitive advantage in a global market.

See real time voice translation in this two-minute video. Learn more at Cognigy.com or visit us on the Genesys AppFoundry® Marketplace.