From Hype to Impact: Why Operational Excellence Unlocks AI's Value

Why AI initiatives stumble

AI discussion dominates boardrooms and headlines, yet many implementations fail to deliver tangible returns. A recent MIT study found that only 5% of generative AI pilots produce measurable profit-and-loss impact, while a Goldman Sachs analysis showed 58% of S&P 500 companies mentioned AI in Q2 earnings calls. The contrast between attention and results points to gaps beyond the models themselves.

Operations: the bridge between promise and practice

Speed matters in AI adoption, but rushing often means skipping the fundamentals that make technology work in a business context. Survey data from Lucid shows more than 60% of knowledge workers think their company’s AI strategy is only somewhat or not at all aligned with operational capabilities. When processes are undocumented, ad-hoc, or fragmented, automation amplifies inefficiency as much as it amplifies efficiency.

The ‘last mile’ of AI

The hardest part of AI transformation is not building models, but embedding them into daily workflows — the so-called last mile. Organizations may have access to powerful models but fail to connect them to the people and documented processes that actually create value. Lucid’s survey found that only 16% of respondents say their workflows are extremely well-documented. The top obstacles to better documentation are lack of time (40%) and lack of tools (30%).

Collaboration and change management are often overlooked

Perception of AI strategy varies by role: 61% of C-suite executives feel the strategy is well-considered, compared with 49% of managers and only 36% of entry-level employees. That mismatch signals a need for structured, collaborative planning and clear decision-making. Even when AI speeds up tasks — for example, generating a preparatory memo in minutes — teams still need forums to debate, prioritize, assign ownership, and record decisions.

What teams actually need

When asked what would help them adapt to AI, respondents prioritized basic operational tools: document collaboration (37%), process documentation (34%), and visual workflows (33%). These choices underline that organizations are not asking for more advanced AI; they need modern collaboration and documentation systems to harness existing AI capabilities.

How to move from pilots to production

Leaders should focus their investment on operational rigor rather than assuming technology alone will deliver results. That means documenting workflows, standardizing processes, equipping teams with collaboration tools, and establishing clear change-management practices to ensure AI outputs are adopted and tracked. In short, the path to realizing AI’s potential runs through operational excellence — especially the last mile of integration into everyday work.

Source and context

This analysis is based on Lucid Software’s AI readiness survey and related industry studies. It highlights practical, often low-cost improvements that can increase the ROI of AI initiatives without relying solely on acquiring top-tier AI talent.