As customer experience leaders move deeper into AI from pilots to commercial-scale implementations, the pressure is shifting from experimentation to evidence. It is no longer enough to run a promising proof of concept. Businesses now need to show how AI improves operational performance, reduces cost, strengthens customer journeys or helps teams manage risk.
In this CX Today interview, Nicole Willing speaks with Anshuman Singh, Chief Executive Officer at HGS Europe, about why so many AI pilots struggle to turn early promise into meaningful business value.
For Singh, the issue is that in many cases, businesses have not defined the right measures of success before they begin.
“It’s less about, you know, that AI failed. It was more about that we couldn’t prove that AI succeeded.”
That point is becoming increasingly important as organizations move from the excitement of generative AI into the practical work of enterprise adoption. Singh notes that large language models made it easier than ever for businesses to launch pilots, particularly around assistive AI. However, easy experimentation can create a false sense of progress if leaders have not agreed what the initiative is supposed to achieve.
In CX, that might mean improving first contact resolution, lowering cost per resolution, increasing customer retention, or supporting risk mitigation in regulated environments. Without a clear target, even a technically impressive pilot can lose momentum.