The Challenge
Many organizations launch generative AI experiments without defining concrete value metrics upfront, leading to pilot fatigue and unclear business returns. Furthermore, technical leaders often assume that building a superior tool guarantees organic user adoption, underestimating the heavy organizational change management and cross-functional buy-in required to make AI part of daily operations.
Core Findings
Zoetis established an enterprise program (AI@Zoetis) paired with a Business Value Realization Framework to guide investments across three distinct tiers: grassroots innovation, functional value, and core enterprise transformation. By rolling out a model-agnostic generative AI platform granting access to 10 large language models, Zoetis achieved a 95% employee adoption rate and generated over 11,000 colleague-built agents. In customer experience, AI-driven real-time interaction analysis expanded customer feedback signals tenfold compared to traditional NPS polling. CDTO Keith Sarbaugh emphasized that partnering closely with the Chief Human Resources Officer on workforce upskilling and intentional change management was essential to turn early experimental bets into scaled enterprise victories.
Strategic Takeaway
Before launching technical proofs of concept, define exact value metrics across defined categories such as revenue uplift, cost reduction, or real-time workflow intelligence. Democratize access to safe, multi-model platforms to encourage bottom-up agent development, while reserving major capital for high-conviction functional bets. Most importantly, treat organizational change management as a core technical discipline by partnering directly with HR leadership.