01

What is changing or failing?

Technology pilots often begin with a promising capability and willing sponsor but an imprecise operating problem. The team demonstrates that the technology can work while leaving unanswered whether it changes a valuable outcome in the refinery’s actual environment.

When the pilot ends, data limitations, integration effort, cyber review, process ownership and adoption requirements appear as new questions. The organisation has learned about the tool but not made the scale decision easier.

EY’s 2025 Future of Energy Survey illustrates the adoption-to-value gap. It reports that 50% of oil and gas and chemicals respondents were using digital twins, while only 14% of users said the technology was living up to expectations. The sample covered 508 oil and gas and chemicals respondents within a 1,020-person global energy survey. This is a perception survey, not an ROI benchmark, but it is a useful warning: deployment and value realisation are different management outcomes.

02

Why does it matter commercially?

The direct pilot cost may be modest. The larger cost is accumulated management attention, integration work and a portfolio of initiatives that cannot be stopped or scaled with confidence. A demonstration can be technically successful while leaving the refinery unable to name the budget line, operating owner or measurable benefit.

A weak pilot also distorts vendor selection: presentation quality and local enthusiasm can outweigh evidence of operational fit, economics and implementation readiness.

03

What must management decide?

Before launch, management must define the decision the pilot will support: proceed to scale, reshape the use case or stop. Five conditions should be explicit—problem, baseline, benefit owner, evidence threshold and scale gate.

Data access, security, technical authority and operating adoption remain conditions of the decision, not details to be solved after success is declared.

Require a one-page scale decision

The final pilot output should state: baseline, measured result, confidence and limitations, recurring operating cost, implementation dependencies, named benefit owner and recommendation to scale, reshape or stop. A technical demonstration without this decision record is unfinished work.

04

What evidence is required?

  • Named operating problem and measurable baseline
  • Benefit owner accountable beyond technology delivery
  • Relevant evidence for the site, workflow and operating context
  • Data, integration, cyber and human-workflow dependencies
  • Pre-agreed acceptance threshold and scale decision date
  • Named budget line and cost the use case is expected to remove or protect
  • Recurring run cost, adoption effort and benefit sensitivity at full operating scale
  • Explicit evidence that would cause management to stop rather than extend the pilot

05

What should happen next?

  1. Rewrite the pilot charter around one management decision, one operating owner and one measurable value line—not the product demonstration.
  2. Separate vendor claims from evidence and open assumptions.
  3. Name the operating owner, control authorities and adoption responsibilities.
  4. Start with the smallest read-only evidence set that can test the value case.
  5. Do not increase integration effort or rollout capital until the scale conditions are met.