Only 11% of operations leaders surveyed say their operations technology investments have delivered expected results. Sixty percent are scaling AI across the enterprise, but only 31% report broad business impact from their investments. In PwC’s 2026 Digital Trends in Operations survey of 767 US operations leaders, ambition is clearly ahead of results.
Many operations AI programs stall between the pilot and the profit and loss (P&L) statement. Companies run diagnostics to find opportunities that may not get funded, design future states nobody tests, and roll out technology without the operating model changes it depends on. Each step starts from scratch, and the client pays for the restart. Leaders cite integration complexity (52%), data issues (46%), and user adoption (43%) as the top reasons investments fall short. Trust is the other gap. Leaders expect agents and automation to handle operational tasks, yet only 37% are comfortable letting agents run a process holistically.
The result is that boards are funding fewer pilots and asking for proof before capital is committed. For COOs, CSCOs, and CPOs under pressure to show AI returns in operations, the question is no longer where AI can be used. It is what they would need to prove before committing capital, and how quickly they can prove it.
Many operations AI programs stall because each step starts over. Operations Intelligence helps connect them. Its three modules run as one sequence on one technical foundation, so what is learned in each stage carries into the next.
Diagnose finds and sizes the value, including trapped cost and working capital, constrained growth, and the operational levers behind them. Leverage frontier LLM and agents to help accelerate and go deeper than previously possible. The business case starts from measured evidence rather than assertion.
Innovate designs the future operating model and tests it in a digital twin against real constraints and volatility. Proprietary AI harnesses, agents, and templates help quickly build near-production applications. Leadership can see the model work, or see where it breaks, while change is still cheap, helping to build trust.
Deliver executes the change with agents working inside the implementation rather than alongside it. Agents support both the work itself, but also orchestration and traceability across the lifecycle of complex tech deployments. This helps shorten timelines and reduce the risks and cost of delivery.
Running across all three is a knowledge layer that carries client context, sector and sub-sector context, and PwC’s proprietary operations IP. Each stage starts from what the client and the industry already know, and the knowledge sharpens with each engagement.
Many organizations struggle to prove returns before committing capital, and programs often stall before they show results. PwC brings decades of transformation experience to help clients put proof first. With Operations Intelligence, diagnostic findings shape the design, the design is tested before money is spent, and the delivery team works with a business case that has already been stress-tested.
Clients get shorter implementations and faster time to value when this approach carries into delivery. In early implementations, agents working inside the Operations Intelligence delivery process have produced a ~60% productivity gain. Clients can also see the approach running in a comparable context before committing.
The useful starting question is not “where can we use AI?” It is “what would we need to prove before we commit capital, and how fast can we get there?” Companies are funding fewer pilots, so each one should show returns sooner. PwC works alongside clients to answer that question, using Operations Intelligence to build a sequence leaders can follow, with evidence at every step.
Explore PwC's solutions for business and supply chain operations transformation. Explore now.
Discover how PwC helps operations leaders move beyond disconnected AI pilots, accelerate transformation, and deliver measurable business value through intelligent, connected operations.