In this episode of Tech Talks Daily, I spoke with Keith Zubchevich, CEO of Conviva, to explain one of the most honest analogies I’ve heard about the current AI rollout.
Keith likens modern AI agents to toddlers sent out to get a job, full of promise, curious and energetic, but still lacking the judgment and context needed to operate safely in the real world. It’s a simple metaphor, but it captures the tension many leaders are feeling as generative AI matures in theory, while so many implementations fail in practice.
As ChatGPT approaches its third anniversary, the story suggests that GenAI has come of age. Yet Keith argues that this sense of maturity is misleading, especially within companies chasing measurable returns. He explains why so many pilots get stuck or quietly disappoint, not because the models lack intelligence, but because organizations often release agents without clear outcomes, real-time monitoring, or insight into how customers actually experience those interactions.
The result is AI that seems to function well internally, while silently frustrating users or failing to complete the task they were intended to perform.
We also delve into the now infamous Chevrolet chatbot incident where a $76,000 truck was sold for one dollar, and use this as a lens to explore what happens when agents are left without boundaries or oversight.
Keith makes a strong case that the next chapter of enterprise AI will not be defined by ever-larger models, but by visibility. He shares why observing behavior, patterns, sentiment, and efficiency in real time is more important than chasing pure accuracy, especially once AI moves from internal workflows to customer-facing roles.
This conversation will resonate with anyone under pressure to rapidly scale AI and concerned about brand risk, responsibility, and trust. Keith offers an informed view of what effective AI “parenting” looks like within modern organizations, and why measuring customer experience remains the most reliable signal of whether an AI system is actually maturing or simply creating new problems quickly.
As leaders rush to put agents into production, are we really ready to guide them, or are we sending toddlers into the workforce and hoping for the best?
Useful links
Subscribe to the Tech Talks daily podcast
#Conviva #CEO #turning #experimental #agents #reliable #systems