Templates lower the cost of getting started

What it is

To write a competitor comparison, someone can invent structure from an empty box, or open “comparison table” and find columns, an object slot, and a tone slot already laid out. Template startup scaffolding means the stretch of work that compiles intent into a prompt is shared by a structure built in advance, so time to a first submittable draft drops. It solves “zero to first sentence,” not “what can this product do.” Discovery is an example’s job; a template assumes a task has already been chosen.

What drops is compilation cost, not generation quality. The submission can still be bad. They are simply no longer stuck at a blinking caret.

Why it happens

Writing and prompting share a start friction: a blank page forces deciding what to say and how to organize at once. A template pins organization first; working memory only fills slots. Slots also act as a checklist: an unfilled “comparison dimension” is a visible gap, and an empty box will not remind you that gap exists. Starting turns from “invent a prompt” into “answer a few known questions.” The cognitive move shifts from generation to recognition, and recognition is faster.

For the model there is an engineering bonus: slots write constraints as stable fields, easier to check downstream. Start cost falls partly because people think less, partly because the system guesses less.

Studying it

One task, empty box versus optional template. Record time to first submit, deletions before submit, whether the first draft contains required slots (object, dimensions, output form). Independent variables: whether the template is optional, slot count, familiarity of the task with the person’s actual job. Dependent variables: start latency, first-draft completeness, subjective “I know what to do next.”

Do not use final quality as the primary metric. A template can lock someone in and the model can still raise or sink the final; this entry only owns the start. Experts who already know the task are a control: their start cost is already low, and the template’s marginal gain should be smaller.

Where it stops holding

When the task has no stable structure — chat, brainstorm, exploring unknown data — there is no slot to pin. Start cost will not fall, and choosing a template adds overhead. Experts who arrive with their own prompt library treat an official template as interference. Once slots exceed what can be filled on the spot, start cost rises again — a form can be more off-putting than a blank. This entry does not treat whether templates lock people inside the designer’s idea of the task.

Applying it

• After a task is chosen and before the empty box, offer one or two templates for that task, slots expanded by default, not a full sample essay.
• Put in slots only what the start actually lacks (object, output form, key constraints). Hide polish options.
• Allow “no template, I’ll just speak” and remember the choice, so people do not pay a template tax every time.
• Check: people who cannot write prompts but can do the task, from opening the entry to first submit. With a template this should be clearly shorter than the empty box, and the first draft should contain slots the empty-box group often omits. If latency does not drop, or it drops but required slots are still missing, the template is only an extra button.

Related

• Same group: L2.03.2 Templates constrain the range of what users will express · L2.03.3 Templates must be editable, not only fillable · L2.03.4 Templates turn open input back into structured blanks, at the cost of openness · L2.03.5 Placeholder content in templates is kept verbatim; empty slots must be unskippable · L2.03.6 Templates encode the designer’s task assumptions and actively mislead when those do not match real work · L2.03.7 A template that can only be applied whole, with its structure invisible, teaches nothing
• Nearby: L2.01 Openness of Natural-Language Commands and Its Cost · L2.02 Discoverability of What Can Be Said · L2.04 Complementarity of Parameter Controls and Natural Language
• Search terms: template startup scaffolding · blank page problem · slot-filling prompt