---
id: con-solution-ladder
schema: concept/1
layer: garden
status: emerging
name: Solution ladder
aka: ["lowest rung that works", "right-sizing the tool", "paper before app"]
summary: >
  A way to choose how to solve a problem by climbing from the simplest option (paper, a document or
  sheet, a prompt) toward an agent or an app, and stopping at the lowest rung that works.
first_seen: unknown
relations:
  builds_on: [con-smallest-useful-version]
  evolved_from: []
  related: [con-agents, con-automation]
  parallels:
    - idea: "paper prototyping"
      field: "design"
      note: "Sketch the screens on paper and test them with people before anyone builds anything."
    - idea: "minimum effective dose"
      field: "medicine"
      note: "Use the smallest intervention that achieves the result; more adds risk without adding benefit."
    - idea: "three kinds of agent: fixed steps, a reasoning agent, several agents working together"
      field: "AI product management"
      note: "Hamza Farooq and Jaya Rajwani advise starting with fixed step-by-step automation for the quickest measurable return (res-other-lenny-ai-agent-categories)."
    - idea: "moving to a bigger tool only when you outgrow the last one"
      field: "nontechnical building"
      note: "Zevi Arnovitz moved from a chat project to Bolt to Cursor only when each stopped being enough (res-other-lenny-nontechnical-pm-cursor)."
implementations: []
competencies: [cmp-frame-the-outcome, cmp-make-it-repeatable]
origin:
  kind: ai-proposed
  by: "seed agent 2026-10-16"
  rationale: "Part of the released Build Cycle method; keeps nontechnical builders from reaching for an app when a prompt would do."
---

When AI can build almost anything, you still have to decide what to build. The solution ladder helps you decide. Picture the options as rungs, from simplest to most complex:

1. Paper or a conversation.
2. A simple document or spreadsheet.
3. A prompt, saved and reused.
4. An agent or automation that runs steps for you.
5. An app other people use.

Start at the bottom and climb only when the rung below can't do the job. Each step up costs more to build, check, maintain and hand over.

The ladder also works in reverse. If a project has grown painful, ask whether stepping down a rung would serve people just as well.

## Example

A team wants "an app to track event sponsorship." Climbing the ladder, they find a shared spreadsheet plus a saved prompt that drafts follow-up emails covers everything they need. They revisit in six months; they still haven't needed the app.

## How it connects

- Builds on the **smallest useful version**: decide what's needed, then the lowest rung that delivers it.
- Higher rungs are **automation**, **agents** and **vibe coding**.
- Choosing the rung is part of **steering** the whole project.

## Try it

Think of a tool you wish you had and write down which rung you think it needs. Then describe the problem to AI and ask it to argue for the lowest rung that could work.