编写设计系统立项论证方案
依据组织现有证据比较建设选项,整理成本收益、试点指标与资源申请,供管理层决策。
完整提示词
<context>
Leaders fund outcomes, not components. Design system pitches fail when they lead with consistency and craft, quote industry savings percentages as if they applied here, or hide the ongoing cost: a system is a product that needs a team after launch, and product teams pay an adoption cost too. Persuasive cases use the organisation's own evidence (the number of date pickers in the codebase, hours spent rebuilding the same UI, accessibility defects and the effort to fix them, how long a new screen takes), turn it into ranges with stated assumptions, compare real options including "do nothing" and "adopt an open-source base with a thin layer", and propose a pilot with a checkpoint so leadership can stop if the numbers do not appear. Accessibility obligations vary by country and sector, so the case should state them as something to confirm with legal, not as a threat.
</context>
<task>
Build the business case.
<situation>
{{situation}}
</situation>
The ask: {{ask}}
{{#evidence}}
<evidence>
{{evidence}}
</evidence>
{{/evidence}}
1. **Problem.** Describe the cost of the current state in the organisation's terms (slower delivery, duplicated work, inconsistent experience, accessibility defects, onboarding time), backed by the evidence given. Where evidence is missing, list what to collect in two weeks and how.
2. **Options.** At least three: do nothing, a minimal option (for example adopt an open-source component base plus tokens with part-time owners), and the requested option. For each, what it delivers and what it does not.
3. **Costs.** People, tools and the adoption effort product teams will spend, for year one and ongoing. Use the user's cost figures; if none are given, use placeholders such as "[loaded cost per engineer]" rather than inventing salaries.
4. **Benefits.** Estimate time saved and defects avoided as low, expected and high ranges, each with its formula and assumptions, starting from the organisation's own numbers. Separate hard savings from softer benefits (brand consistency, hiring appeal).
5. **Assumptions.** Every number's source, labelled measured, estimated or placeholder.
6. **Risks.** Low adoption, the team becoming a bottleneck, scope creep, and dependence on a few people, each with a mitigation.
7. **Measures of success.** Three to five measures with a baseline and a target at 6 and 12 months, and a go or stop checkpoint for a pilot.
8. **The ask.** One paragraph: what, how much, for how long, and what leadership gets back by when.
9. Write for a reader with five minutes: the summary must stand on its own.
</task>
<constraints>
- Do not invent statistics, studies, salaries or company examples. If you mention published industry figures, say they vary widely and are not evidence for this organisation.
- Do not claim specific legal requirements; say accessibility obligations depend on jurisdiction and sector and should be confirmed.
- Keep benefit estimates conservative and show the arithmetic.
{{> output/uncertainty}}
</constraints>
<output_format>
Markdown with the sections from the contract, in order. Summary under 150 words. Options, costs and benefits as tables with formulas visible. End with Open questions.
</output_format>可替换内容
- {{situation}}
- {{ask}}
- {{#evidence}}
- {{evidence}}
- {{/evidence}}
- {{> output/uncertainty}}
- [loaded cost per engineer]
分类与标签
办公效率、立项论证、设计系统、成本收益、管理层汇报