AI for the Telecoms Contact Center: Think Marathon, not Sprint

By Matt McConnell, CEO of Intradiem

The current incarnation of AI is the biggest tech breakthrough since the Internet. So far, we’ve only seen the tip of the iceberg.

Big data and expanded computing power have paved the way for a new artificial intelligence (AI) whose models can digest the universe of online information and help us solve nearly any identifiable challenge. Customer service is an ideal place to exploit this powerful new capability. Despite heavy reliance on automation, contact centers still suffer from inefficiency and a reputation for poor customer experiences.

AI can change that—not by taking over, but by reinforcing. We all complain about inefficiency, but who wouldn’t agree that the only thing worse than working with a less-than-helpful agent is not being able to reach one at all? Especially when that customer is dealing with internet outages or trying to set up their router or modem in a new home. No technology can substitute for the innate compassion and judgment of a live agent. That’s why we need to assess AI according to its ability to help those agents become better customer servants.

First, Proceed with Caution

Service providers exploring ways to use AI need to be intentional as they map out plans to integrate it into their contact centers. It’s still early, and AI is advancing in fits and starts. Some companies are actively probing its limits, but most are either waiting on the sidelines or dipping just a toe or two into the experimental waters. The roadmap that eventually emerges must be based on a thorough assessment of opportunities and challenges, and that will take time.

One significant challenge is the fact that AI models require consistent data. Today’s telecoms industry is the product of multiple mergers between companies that stored data in a variety of formats. Any AI initiative would need to begin by standardizing those formats. Another challenge is AI’s poor track record on accuracy and security.

This poses a dilemma for the development and commercial exploitation of AI. Aligning potential with reality and calming justifiable anxiety will require AI providers to demonstrate greater usability, security, and effectiveness of their technology. That will require training their models on more data, which will heighten concerns about misuse. Given the financial stakes, any attempt to place privacy restrictions in the path of profitable data exploitation at this point may already be unfeasible.

Then, Focus on Agent and Customer Experiences

As I said at the outset, the correct use for AI is to help make human agents better customer servants rather than to replace them altogether. We’ve seen big disappointments in the past when new technologies failed to live up to expectations. First-generation chatbots and offshore customer service centers seemed like great ideas, for example, but their limitations resulted in customer experience issues and brand damage.

Improving the agent experience should be the first priority. Rigid schedules, repetitive work, and regular flak from frustrated callers make the agent’s job famously stressful. That fuels chronic attrition, which is expensive and destructive.

AI will turn that around by offloading repetitive tasks and simple inquiries, and providing stronger support to help agents deal successfully with the high-stakes interactions they’ll spend more of their time focusing on. The frontline agent’s job will become increasingly professionalized, requiring specialization and additional training. That should lead to higher pay and greater satisfaction.

AI will also improve the customer experience. For example, there are many telecom service providers that are currently testing AI’s ability to:

• Predict “next best actions” by analyzing customer interactions and using those insights to facilitate agents’ ability to resolve subsequent calls that involve similar situations.
• Preempt customer service problems by predicting when households or service areas are approaching capacity-limit thresholds and triggering alerts and automated corrective action.
• Accelerate customer service calls through smarter diagnostic capabilities that can identify causes of problems in real time and shorten the time needed to fix them.

If your business’s contact center is ready to level-up operations though AI adoptions, use this four-step framework as your guide:

• Understand its capabilities. AI can enhance speed and efficiency, but it has little flexibility and no empathy. Your customers know the difference.
• Know what you want to solve. Reach beyond a vague fear of falling behind competitors and carefully define real problems you need to solve.
• Target the best opportunities. Standardized data is essential to training AI. Identify areas of opportunity in your business that are both simple enough for AI to handle and where uniform data exists to train the AI.
• Start with low-risk initiatives. Don’t experiment on your customers. Start with internal-facing AI projects and use your experienced agents to vet your AI technology to ensure the accuracy and effectiveness of its outputs.

With careful preparation, new AI technology will streamline efficiency chokepoints like scheduling, training, in-call support, call-handle time, and administrative follow-up work. AI will play a transformative role, but we should never forget that customer service is an exchange between human employees and human customers.

About Matt McConnell, CEO of Intradiem

Matt is Chairman and CEO of Intradiem. He founded the company in 1995 with a vision of reinventing customer service through automation and artificial intelligence. Today, Intradiem is the leading provider of Intelligent Automation solutions for customer service teams. Matt graduated from The Georgia Institute of Technology with a Bachelor of Science degree in Industrial and Systems Engineering.

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!

Categories: Artificial Intelligence · Industry Viewpoint · VoIP