Can AI Automate Live Chat? Dr. Michael Housman Speaks to General Assembly

In a session with General Assembly, Dr. Michael Housman discusses how organizations should think about deploying artificial intelligence for live chat, emphasizing a pragmatic, human-centered approach rather than full automation.

Housman explains that live chat is a uniquely challenging communication channel. Unlike face-to-face or voice interactions, chat lacks non-verbal cues such as tone, facial expression, and body language. As a result, small linguistic choices—wording, timing, punctuation, and emotional signals—carry disproportionate weight in shaping customer perceptions and outcomes.

Rather than positioning AI as a replacement for human agents, Housman advocates for a “human-in-the-loop” model. In this approach, AI systems analyze large volumes of historical chat data to identify conversational patterns associated with positive outcomes, then surface guidance to human agents in real time or through coaching and feedback.

The value of AI in live chat, he argues, lies in its ability to augment human judgment. Machine learning excels at detecting subtle, non-obvious patterns across millions of interactions, while humans retain the ability to interpret context, exercise empathy, and make nuanced decisions when conversations deviate from the norm.

Housman cautions against premature reliance on chatbots for customer-facing sales or service roles. Current bots struggle to adapt dynamically, personalize responses, or reflect brand identity in a consistent and authentic way. Over-automation can therefore degrade the customer experience rather than improve it.

Instead, he encourages organizations to deploy AI first as a measurement and optimization layer—one that helps companies understand which conversational behaviors drive satisfaction, conversion, retention, and revenue. These insights can then be used to train and support human agents more effectively.

The broader message is that successful AI deployment in live chat is not about removing humans from the loop, but about combining the strengths of machines and people. When AI is used to inform, guide, and learn from human interactions, it can meaningfully improve both customer experience and business outcomes.