Cure The Automation Hangover In Customer Service
Close your eyes and I’ll set the scene: It’s 2016. Chatbots are all the rage. News articles are predicting the death of the contact center as we know it. AI, the prophesied hero, powering the agent who doesn’t need breaks and can engage with 100 — no, 1000 — customers at once. Infinitely scalable human-like intelligence. The dream sequence ends. Come back to reality, and open your eyes.
Five years later, and a few things have been made clear:
- Customers don’t like chatbots nearly as much as we thought they would.
- The capacity for AI to replicate human intelligence has been grossly overstated.
- Human agents are increasingly important in today’s customer service strategy.
The good news? We’re seeing a shift in perspective. Our language is shifting from automation to augmentation. We’re acknowledging that customers are looking to speak with a human agent and that this human needs to be empowered to help them. The bad news? The dream of a human-less contact center has stunted innovation in solutions for that very same agent.
We’re suffering from a wicked automation hangover.
There is a certain irony here: As you increase the complexity of automation, the involvement of human participants becomes more critical, not less. With chatbots managing all of the simple inquiries, agents are left to handle the most complex and emotionally charged conversations. Agent desktops are increasingly infused with AI capabilities, but these solutions are seldom designed with the agent experience in mind and can hinder more than they help. On the back end, AI systems require human oversight for training and optimization, an increasingly specialized skillset. This wouldn’t be as problematic, though, if companies hadn’t underinvested in the very people required to facilitate the vision of an AI-powered contact center.
So What’s The Cure For The Automation Hangover?
Contact centers have realistically automated everything that’s automatable without requiring reorientation from a technology-centric view of agent augmentation to a human-centric one. The next area of differentiation for the contact center won’t lie within any specific technological advancement but rather in how companies enable partnership between agents and AI — not to automate, but to elevate human capability beyond what’s possible alone.
- Stop trying to make your AI more human. AI’s value isn’t in how well it can replicate human intelligence — it’s actually pretty miserable at that. Its value is the way it can amplify human intelligence. Humans and machines are good at different things, but their strengths are complementary.
- Map out your agent journeys. Understanding your agents’ workflow end-to-end will help you identify and prioritize areas for improvement. Consider which tasks are best suited to machine or human partners and how and when they should collaborate. But don’t fall into old habits: Apply human-centered design methods to ensure any augmentation solution delivers the desired value.
- Trust your humans. More specifically, trust that your humans know how to human better than your tech stack. AI is lousy at understanding novel situations, but your agents aren’t. Empower your agents to ignore AI-powered recommendations if they have a better solution.
- Start building the flywheel. Closing the loop on when an agent took AI-powered recommendations (and just as importantly, when they didn’t) will help identify areas of improvement for your algorithms, both to reinforce current predictions or to correct where needed.
AI is an incredible tool that can deliver enormous value in the contact center, but it needs to be leveraged intentionally. It’s time to approach human problems through a human-centered lens.
Having the opportunity to dig into these topics (and others!) is one of the main reasons I joined Forrester. My research will be focused on solutions and strategies that will help customer service and customer experience (CX) leaders transform agent-assisted interactions in the contact center.
Are you working on solving these problems? Got feedback? I’d love to hear from you. Get in touch!