A pilot needs somewhere to go.

A team can spend months exploring a technology without anyone being responsible for what happens afterwards. We propose starting every potential pilot discussion with the intended decision: who will review the result, who can authorise continuation and which alternatives they will compare. This does not require a promise to buy. It requires a credible explanation of how the evidence could inform a real organisational decision rather than become an interesting report with no next meeting.

A useful operational result may still leave unresolved assurance, budget, data access or support requirements. NHS DTAC provides a buyer-assurance reference for relevant digital health technologies (NHS England, n.d.). The actual requirements would depend on the intended use and organisation. We propose mapping these conditions early enough to avoid an evaluation whose positive result cannot be acted on. An assurance discussion is part of planning, not proof that PHARMORIS is approved, procured or ready for a particular setting.

The evaluation should account for preparation, checking, training and follow-through under realistic conditions. A time saving in one activity may be offset by additional work elsewhere. The Magenta Book provides evaluation-design context; the appropriate method would be chosen for the actual question (HM Treasury, n.d.). We want a comparison a buyer can interrogate, including the baseline's strengths and the uncertainty around the result. That is more useful commercially than a favourable headline the partner cannot defend internally.

A well-designed pilot can end because the benefit is too small, the support burden is too high or the organisation cannot use the result. Those outcomes protect both parties from a larger unsuitable commitment. We propose agreeing stopping conditions before the pilot begins, including consequential output errors, unusable data and an unfair comparator. A partnership should not be trapped by the sunk cost of a demonstration or pressured to provide a testimonial that the findings do not justify.

The investment discussion is about financing bounded learning with a commercial destination: a qualified buyer, a meaningful evaluation and a clear decision afterwards. We would expand only when the result supports further work and the delivery conditions are viable. We would redirect if pilots repeatedly produce interest without purchasing authority or an adoption route. No pilot partner, contract or conversion rate is claimed here. The thesis is a disciplined way to earn those milestones, rather than forecast them as inevitable.