Empathy AI: Top Analyst Calls on Tech Companies to Focus on Solving Real User Problems
Erick Brethenoux, Gartner’s global chief of AI research, is advocating for a paradigm shift in the implementation of artificial intelligence. His belief is clear: AI should prioritize simplifying users’ lives by automating mundane tasks and addressing genuine problems, rather than merely generating meeting summaries or focusing on less impactful applications. Brethenoux brands this vision “Empathy AI,” emphasizing technology’s potential to empathetically meet user needs and eliminate dreary tasks.
During the Gartner Data & Analytics Summit in Sydney, Australia, Brethenoux voiced his critique of the status quo, asserting, “AI is not doing its job today and should leave us alone.” His comments emerge in a climate saturated with excitement over generative AI and AI agents—fields he contends are often exaggerated by technology vendors.
A central issue Brethenoux highlighted was the impracticality of AI-generated meeting summaries. He mentioned that users rarely have time to review summaries from years past or follow up on their suggested action items. Instead, he champions a model where AI improves our daily experiences by autonomously tackling monotonous tasks, thus leveraging what he terms “Empathy AI.”
Brethenoux illustrated his point with successful case studies, such as the US healthcare company Vizient and a real estate firm, both of which adeptly implemented Empathy AI. These organizations pinpointed employees’ least favorite tasks and seamlessly automated them with AI, leading to instant acceptance and no resistance in change management.
However, Brethenoux voiced caution regarding the prevailing trend of tech vendors promoting AI agents for automation in intricate enterprise scenarios. While AI agents have functioned in relatively controlled industrial systems for years, Brethenoux noted their challenges when facing highly complex tasks.
He questioned the practicality of personal AI agents collaborating across various enterprise data sources to autonomously make nuanced decisions, like scheduling meetings. Addressing how such systems would reconcile competing priorities among an employee’s boss, partner, or kids, Brethenoux noted the usual vendor response was marked by “silence.”
To build effective agentic systems that address these complexities, Brethenoux underscored it as a profound software engineering obstacle. This challenge necessitates well-thought consideration regarding system decomposition, communication, autonomy levels, and information management and control.
Furthermore, Brethenoux pointed out the conflated usage of terms such as “AI agent” and “generative AI” by vendors, contributing to the overall AI hype. He echoed an adage from French philosopher Albert Camus: “To misname things is to contribute to the world’s miseries,” stressing the necessity for precise terminology in the AI domain.
