Four Companies, One AI Playbook: It's a Leadership Problem, Not a Tooling Purchase

McKinsey’s Jeff Hart interviewed four enterprise executives on how they’re operationalizing AI at scale, and the pattern across Dell, Mercedes-Benz, AMD, and Liquid AI is the same regardless of industry: AI adoption succeeds or fails on leadership and change management, not on which tool gets purchased. Dell’s chief strategy officer Sam Burd describes decoupling revenue growth from cost growth while improving customer outcomes — something he says required company-wide commitment, not isolated pilots, because value capture ultimately comes down to leadership commitment and employee adoption. AMD’s chief administration officer Ruth Cotter details a mandatory, company-wide AI training program meant to establish baseline proficiency and signal that AI fluency is required in every function, not just engineering. Mercedes-Benz’s Magnus Östberg is building software-defined cars around conversational AI, and Liquid AI’s Ramin Hasani is betting on edge deployment across billions of existing devices — different problems, same structural conclusion: sequencing and adoption discipline matter more than the model.

That same convergence is showing up independently in B2B GTM circles. Three separate SaaStr AI 2026 sessions — Atlassian, Anthropic, and Scale Venture Partners — landed on an identical “both/and” playbook without coordinating: Anthropic’s Eleanor Dorfman said threading Claude through existing sales tools rather than replacing them pushed self-serve enterprise adoption to 54% of new logos and returned 70% of AE time from admin work to selling, while Scale’s Rory O’Driscoll argued the real moat isn’t the model but the skill/harness layer built on top of it.

Seven executives across two unrelated conversations, one shared instruction for anyone rolling out AI on a commercial team: the tool decision is the easy part. The sequencing — who gets trained first, what existing workflow the AI slots into, who owns adoption — is the part that actually determines whether the rollout produces the 70%-time-back outcome or another shelved pilot.