Most AI rollouts don't fail because the technology is wrong. They fail because no one agrees on who owns it or what success looks like.

That is the clearest signal in Clarinet's State of AI in the People Function, a survey of 208 People and HR leaders fielded in June 2026, set alongside Clarinet's adoption work with more than 70 organizations. Seventy percent of leaders named a place where AI fell short of what they had hoped for. Among those who gave a specific reason, the largest was not model quality and not integration. It was readiness: skills, training, and strategy.

Diane Sadowski-Joseph, Co-Founder and Head of Product at Clarinet, walks People and HR leaders through what separates organizations making real progress from those stuck in pilot mode, and why the sticking points are almost always human rather than technical.

What the research surfaced:

• 70% saw AI fall short somewhere, and readiness was the leading reason, not the model
• Only 1 in 8 say their People function is leading the rest of the organization on AI
• Job loss ranked sixth among worries at 6%, well behind data security and simply keeping pace
• Roughly 3 in 4 describe an AI policy that is not fully working
• 1 in 6 dropped an AI tool in the last six months, and most just switched to another one

You will leave able to benchmark your own adoption against real data, define clear ownership across HR, IT, and emerging AI roles, diagnose where resistance is actually coming from, and communicate a rollout that addresses job-security concerns head on instead of letting them fester.