Step 7 — Scale
Projected Impact
These numbers are hypothetical projections for this prototype, not measured results.
What is measured, and what is projected
Measured in this prototype: 214 fields reduced to 8 questions, Grade 13.2 wording rewritten to Grade 6–8, accommodations moved from page 14 to the first message, and every screen usable by keyboard with adjustable text and contrast. Everything below is a projection with the assumption stated, so you can judge the number instead of trusting it.
60%
reduction in form abandonment
Projected impact
Assumption: Assumes progressive disclosure, no timeout and save-and-resume remove the most common reasons people stop.
50%
faster completion
Projected impact
Assumption: Assumes 214 fields collapse to 8 questions, with repeats removed and answers reused.
70%
fewer accessibility-related support requests
Projected impact
Assumption: Assumes accommodations offered up front prevent the calls that ask where the help is.
100+
languages possible
Projected impact
Assumption: Based on the languages large language models already handle; each needs human review before launch.
24/7
Agentforce assistance
Projected impact
Assumption: Availability of the assistant itself; a human caseworker still keeps office hours.
From compliance to inclusion
Compliance asks: did we pass?
A checklist is measured once, against a static page, and it can be satisfied by a document nobody can actually finish. Passing an audit is not the same as being able to apply for help.
Inclusion asks: could this person finish?
AI lets an experience adapt in the moment — reading level, language, pacing, input method, and support — so one experience can fit many different people instead of averaging them all.