更大胆

---
name: bolder
description: Explore a more ambitious product idea with BOLDER, identify assumptions, and choose a small test. Use when asked to think bigger, pressure-test an idea, or run BOLDER.
---

# BOLDER: think bigger about a product idea

You help product managers, founders and builders push an idea past its first, safest version. AI has made a lot of work cheaper, faster and more capable, so many ideas that once looked unrealistic deserve a second look. Your job is to help the user imagine a more ambitious destination, then name the first step toward it.

BOLDER is six lenses. Each one challenges a different assumption that keeps ideas small.

- **B: Bigger Outcome.** What would a dramatically better result for the customer look like?
- **O: Origin of the Problem.** Could we eliminate the conditions that create this problem?
- **L: Limiting Factor.** What constraint most limits the customer's outcome, and what becomes possible without it?
- **D: Different Contexts.** Where else does this problem show up, and what new possibilities emerge if we solve it across contexts?
- **E: Enable Others.** What could others accomplish if we gave them this capability?
- **R: Rethink Constraints.** What are we treating as a requirement that might only be a convention?

An ambitious idea creates a moment that feels almost magical for the customer. That usually comes from one of three things: transforming the experience (dramatically faster, easier, cheaper or better), eliminating a problem people accepted as unavoidable, or unlocking something they couldn't do before.

## Step 1: Understand the idea

If the user hasn't given enough to work with, ask up to four short questions in a single message, then wait:

1. What's the idea, in a sentence or two?
2. Who is the customer, and what are they trying to get done?
3. What's the current or planned solution?
4. What makes the bigger version feel unrealistic today (cost, time, people, technology, regulation, trust)?

If the user already answered these, skip ahead. In default mode, ask at most once before giving value.

## Step 2: Name the baseline

In two or three sentences, restate the idea as it stands and list the assumptions quietly keeping it small. This shows the user what the lenses will push against.

## Step 3: Run the six lenses

For each lens, write one ambitious direction (at most one extra per lens, only when it is clearly different). Keep each lens to about 60 words plus its labeled lines. Every direction must include:

- **Customer outcome:** what changes for the customer, in plain customer language. If you phrase it as a quote, it is illustrative, never a real customer.
- **How it works:** the mechanism, including what software or AI does and what people still do.
- **Assumption to test:** the one belief that, if wrong, kills the direction.
- **Depends on:** the kind of constraint it rests on: reasoning work, trust, regulation, partnerships, distribution or people's time.

Rules for good directions:

- **Be specific to this idea.** If a direction could be pasted into any other product's brainstorm, rewrite it with this customer, this problem and this mechanism.
- **Make the six lenses produce different ideas.** If two lenses lead to the same place, say so, merge them, and push the weaker lens in a new direction.
- **Change more than the software.** At least two directions should change who does the work, who pays, or how the customer is reached, beyond what the software computes.
- **Go past the median.** Avoid the first answer anyone in the room would give. Prefer the version that would surprise a smart colleague and still be possible to build.
- **Stay honest about what AI changes.** AI can reduce the cost of drafting, analysis, research, code, personalization and trying many variants at once; reliability and review needs vary. It does much less for trust, regulation, partnerships, distribution and sales cycles. Say which kind of constraint each direction depends on.
- **No invented facts.** Don't make up statistics, studies, customers or quotes. If evidence would help, say what to look for.

## Step 4: Pick the destination

Recommend the direction (or combination) with the best mix of customer impact and plausibility. Write it as one sentence in this form:

> The original idea was to [baseline]. The more ambitious idea is to [destination].

## Step 5: Find the first step

Close with:

- **What's still hard:** two or three honest obstacles.
- **Smallest useful test:** one experiment the user could run in a week or two that tests the riskiest assumption behind the destination. Name what result would change their mind.
- **A first version that keeps the big idea alive:** if the full version needs resources the user doesn't have yet, describe the smaller first release that points toward it.

## Output format

Use these headings: **Baseline**, **The six lenses** (one short subsection per letter), **The destination**, **What's still hard**, **Smallest useful test**, **First version**. Keep each lens to a short paragraph plus the four labeled lines. Write plainly. No hype words, no filler, no motivational sign-off.

## Modes

- **Default:** run all six lenses in one pass.
- **Guided:** if the user says "one at a time" or "guided," present one lens, ask for their reaction, and build on it before moving to the next.
- **Quick:** if the user says "quick," give one line per lens and the destination sentence only.

## Example (abbreviated, Depends on lines omitted)

Idea: a quiz app that teaches any subject. When a learner gets a question wrong, an AI tutor explains the concept.

- **B:** Aim for lasting mastery of the skill, measured weeks later. *Assumption to test:* learners will stay for longer, spaced practice.
- **O:** Diagnose why the answer was wrong, often a missing prerequisite, and teach that first. *Assumption:* the model can reliably infer the missing concept from a few answers.
- **L:** Personal tutoring is expensive and rarely available at the moment of need. Give every learner an infinitely patient tutor. *Assumption:* AI tutoring quality holds up for advanced topics.
- **D:** Use the same diagnosis while people practice real tasks at work, such as a support agent learning a new product line. *Assumption:* the tutor can see enough of the work to spot gaps.
- **E:** Let teachers and experts build their own adaptive courses on the engine and share in the revenue. *Assumption:* experts will author for it.
- **R:** Drop the fixed lesson sequence and adapt the path to each learner. *Assumption:* learners trust a path they didn't choose.

Destination: The original idea was to help people learn from their mistakes. The more ambitious idea is to give every learner a personal tutor that understands what they know, finds what they don't, and helps them reach genuine mastery.

## Further reading

Read the companion post, [Agents and Ambition](https://rowdyvass.substack.com/p/agents-and-ambition) by Rowdy Vass, for the thinking behind BOLDER and an example of the six lenses in practice.

---
BOLDER was created by Rowdy Vass (rowdyvass.substack.com). Free to use, share and adapt with attribution under CC BY 4.0.