Use different futures because they change the comparison

STR.4 and STR.9 compare the same twelve-month question under three plausible conditions. No probability is assigned.

Future condition	Device-only continuation	Four-day service preparation	Consequence for the present choice
Commoditization continues and customers seek accountable service	Sustains current contracted service but does not supply the sought differentiated contribution.	Can establish whether an attainable paid trial is economically plausible and worthwhile.	Prefer the bounded preparation to those four days of device diagnostics at the Board’s accepted sacrifice; the twelve-month service direction remains open.
Purchasing freezes for the full horizon	Sustains existing contracts without service-development work.	No customer can undertake a paid trial in time to inform this choice.	If established before preparation, continue device-only and retain the four days for diagnostics.
Customer-data reuse is restricted	Continues under its existing permitted uses.	A customer-local service may remain possible if its required use is permitted.	Continue preparation only while an attainable permission determination leaves a useful service option; otherwise prefer device-only. Cross-customer licensing is not a substitute for missing permission.

This is a qualified comparison, not a forecast of three actual events. STR.5 retains different directions and their filters. STR.6 forms complete options within them rather than treating “AI,” “training” and “platform” as competing strategies.

A customer-local service can have an internal delivery option and a provider-assisted option with different support and access dependencies. Keeping a useful description of an unselected possibility is distinct from funding its trial. A new contribution or constraint can change the option space, not merely rescore old members.