The BIG mistake in AI strategy to avoid in 2026

A few days ago, I was stuck in another “bot-loop” on the Amazon website. I made a purchase that arrived with a mangled box and a damaged product inside. After multiple attempts with their chatbot to resolve the situation, I gave up. The bot could not get past the logic of their menu structure to understand my situation, and, ultimately, I had to wait for a human to call me back to resolve the issue. For me, and I assume for most people, this happens 80-90% of the time with AI agents designed to replace a human in a specific customer interaction. It seems this strategy of “replace the human with a bot” may be relevant in many “back office” situations with rather logical requirements, but in the realm of customer contacts, experiences, and relationships, it is a BIG mistake.

In my last article for Qorus, “Are businesses ready for the benefits and opportunities of AI?”, I wrote mainly about the current challenges that AI creates for businesses from the standpoint of organizational models:

• Organizational structures – how teams are organized for reporting and flow of communication

• Management models – how decisions are taken and how resources are allocated, and

• Performance measures – how things are analyzed to monitor progress and determine success or failure

For this article, I will focus on the BIG mistake that most businesses are making with how they integrate AI in their day-to-day operations – especially in interactions with customers.

For the foreseeable future, the majority of AI implementations will bring some effect for humans as end-users and/or for humans as the managers and shareholders of organizations. So, until such time that AI determines us humans as useless, some general effect for humans must be considered in any AI implementation. Unfortunately, what I see today in most companies is an almost blanket strategy of “replace the human with the bot,” which is simply dysfunctional in almost all situations. While a large part of this dysfunctional outcome is probably the result of the strategic motivation behind the AI implementation, the more plausible cause is a lack of understanding of the evolution of the human brain and the science of human decisioning.