As machine learning, predictive analytics, and automation usher in a new era of data-driven experiences, the latest strategic guidance calls for striking the right balance between people and technology. Artificial intelligence (AI) is offloading simple tasks from customer service agents in the contact center, including form completions and compliance checks, but the technology’s true promise lies in its ability to help customers and contact center agents.

Don’t Replace Agents with Chatbots

In a recent Forrester blog post, titled Stop Trying To Replace Your Agents With Chatbots—No, Seriously!, analyst Ian Jacobs notes that overall “customers dislike interacting with chatbots.” According to the post, most customers “prefer to call into a contact center, or even reach out via social media, rather than experience a poor chatbot interaction.”

In a world where customers get easily frustrated with ineffective chatbots and difficulty connecting with a real person, Jacobs stresses in his post that augmenting customer service agents with AI, rather than trying to replace them, is the key to successful implementations in the contact center.

Whether you’ve already deployed solutions or you’re ready to try out chatbots at your organization, Jacobs offers up some sensible steps to take before fully deploying the technology externally to your customers. This includes using AI for internal uses, such as enabling agents to more quickly search knowledge bases with a natural language interface. AI can also help agents achieve better outcomes with automated suggestions and next best actions, as well as improved efficiency with “intermingled workflows” that blend AI-completed tasks with live agent support.

How to Get the Human Part Right

A recent article from the MIT Sloan School of Management also discussed the importance of designing AI systems with humans in mind. The article provides some additional guidance, including five tips for implementing “people-centered” AI solutions for better results. These include:

1. Classify what you’re trying to accomplish with AI: Automation, augmentation, discovery, risk mitigation, or compliance? Knowing what you are trying to achieve before you set out on your AI journey is critical to success.

2. Embrace guiding principles: According to the article, guiding principles include transparency that makes the “high-level details” of AI projects known to all involved. Additionally, explainability that makes it clear how your AI system makes decisions is required, as well as knowing how business intelligence is used.

3. Establish data advocates: Hiring chief data officers is not enough, according to the article. Companies are encouraged to establish stakeholders from across the organization to help ensure data quality, accuracy, and proper ethics in AI initiatives.

4. Practice mindful monitoring: Testing data sets to help prevent algorithm bias reduces risk and ensures the best results. Companies should also regularly assess data for reliability, relevance, and new uses.

5. Ground your expectations: Long-term success requires managing expectations with both internal and external stakeholders. Ask the right questions along the way, including how your AI system impacts society and what the right safeguards are to ensure effective and ethical usage.

For a closer look, you can read the complete MIT Sloan School of Management article here.