Enterprise AI returns lag spending; shift to workflows

The interview with Thomas Robinson points to a fundamental shift in how enterprises should measure AI success. Rather than counting seats or tokens, leaders need to tie AI directly to revenue, innovation, and cost cutting. This is especially critical in sectors like pharma and finance, where errors have high stakes. Robinson's assertion that "judgment is better than hallucination" underscores the limits of current reasoning models and the need for human oversight. His observation that model providers are building forward-deployed engineering teams suggests that the real value lies not in the models themselves but in their integration into business workflows. This implies that the competitive advantage will shift to those who can effectively govern and implement AI in mission-critical processes.