AI Radar is Jasmin Guthmann’s monthly column on Agentic AI, business transformation, and the future of digital value creation. With many years of experience in technology, transformation, and AI, she is one of the established voices shaping the conversation around the future of digital business. In April 2025, her book If There Is a Will, There Is a Way was published — a no-compromise guide to unlocking potential by focusing on what truly matters. Her message: take action, embrace responsibility, and stand for what you believe in.
In the diva-e Conclusion column AI Radar, she shares insights from her practical experience, provides context on current developments, challenges trends, and explores what technological change means for businesses, leadership, and organizations.
The Pilot Trap
95 percent. According to the 2025 MIT study “The GenAI Divide”, that is the share of enterprise AI pilots that fail to deliver measurable business impact — despite $30 to $40 billion in investment.
Spoiler: The models are not the problem. Access is not the problem either. Any company that is serious about AI today has licenses, copilots, and at least one chatbot in its intranet. Getting access to AI was yesterday’s challenge. Creating value from it is today’s.
Welcome to Pilot Purgatory
“Pilot Purgatory” describes the state in which companies move from one successful proof of concept to the next, while hardly any project makes it into productive use. The purgatory of pilots follows a pattern that has become all too familiar:
A use case is identified.
A proof of concept is built.
The demo impresses the executive team.
And then? Nothing. The pilot does not scale because the architecture cannot support it or the data integration is a mess. Because processes built for 2015 do not know what to do with an agent from 2026. Because people who have built twenty years of expertise are reluctant to say out loud: “I don’t know how this works.” That is the real finding. In every second steering committee, people are currently nodding when “agentic workflows” appear on a slide. Nobody asks the difficult questions. Everyone is afraid of falling behind.
It is a bit like The Emperor’s New Clothes: everyone plays along because nobody wants to be the first to say, “I don’t see it.” Except this time, it is not about a garment — it is about billions in AI investments.
There Will Be No Super Agent
The convenient fantasy: one agent that rules them all. A system you buy, switch on, and everything is solved. The reality: a multiverse of agents — specialized, distributed, built by different providers, and powered by interoperable intelligence. Companies that place all their bets on a single ecosystem today are betting against this direction.
Does this mean throwing away the existing tech stack and rebuilding everything from scratch? No. Nobody would get that permission anyway — and that is a good thing. The core can and should remain. But it needs to be reconnected in a way that allows it to evolve with developments we cannot even anticipate today.
This is not a tool decision. It is an architecture decision. And the winners of tomorrow are making those decisions today. The emperor is naked. The question is: who will be the first to say it?
Want to know how ready your organization is for Agentic AI? Start with the Agentic AI Audit.
