# Workflow, Not Wrapper: test an AI product idea, shape the offer and set the price

> A prompt pack and method for anyone building or selling on top of a model. A seven-score absorption test, a one-page workflow product spec, a problem mining loop, a channel translation table, a model column check, an offer ladder with pricing rules, a check for copied success stories, a fidelity gate for generated output, a failure table, a worked example, a rules block and 14 prompts.

- URL: https://paperchasewebb.ai/bots/marketplace/items/workflow-not-wrapper-ai-product-offer-and-pricing-kit/
- Type: Prompt pack
- Category: Solo business
- By: PaperChaseWebb
- Works with: Claude Code, Codex, Claude, Grok Bot, ChatGPT (named in the listing); Plain .md
- Price: $29 USD
- Difficulty: Beginner
- Setup time: 45 min
- Time to result: One idea scored, specced and priced in an afternoon
- Format: Markdown (.md) file, workflow-not-wrapper-ai-product-offer-and-pricing-kit.md
- Version: 1.0, updated 2026-10-27
- License: Personal and commercial use on your own products and your clients' projects. Do not resell or republish this pack.

## What you end up with

You know which part of your idea a model vendor could absorb, what the buyer would still pay you for, and a ladder of prices built on your own cost per job instead of someone else's revenue story.

## What you get

- 13 sections in one .md file
- The seven-score absorption test with a reading guide
- A one-page workflow product spec
- A problem mining loop with a complaint log and grades
- A channel translation table for launch tactics
- The model column check
- A four-rung offer ladder and six pricing rules
- A four-question check for success stories
- A fidelity gate for generated output
- A failure table and a worked example
- A rules block and 14 prompts

## Who it is for

People with an AI product idea or a prototype who want to know whether it is a business or a feature, before they spend months on it.

## Tools needed

- Grok Bot, Claude or ChatGPT: scores ideas, grades complaints and drafts the spec and ladder from your evidence
- A spreadsheet or .md file: holds the complaint log and the absorption scores
- You need: A product or service idea that uses an AI model
- You need: One buyer type you can describe in a sentence
- You need: An hour to read where that buyer talks shop

Cost to run: The pack is the only purchase. Uses the assistant you already have.

## How it works

1. **Score the idea.** Run prompt 1 with the absorption test. Write evidence next to every score. Expected: A total out of 14 and your weakest line.
2. **Mine complaints.** Use prompts 4 and 5 to read where your buyer talks and grade what you find. Expected: A complaint log with Budget-grade entries.
3. **Write the spec.** Use prompt 3 and the one-page spec. Keep the model step to one line. Expected: A workflow product spec with a Not doing list.
4. **Price the ladder.** Estimate cost per job (prompt 10), then draft the ladder (prompt 11). Expected: Four rungs priced on cost and what buyers already pay.
5. **Red team it.** Run prompt 14 and fix the weakest line before you build. Expected: A go, fix or stop decision with reasons.

## Sample prompt: Absorption score (Grok Bot, Claude, ChatGPT)

```
Here is my product idea [paste]. Score it on this seven-line absorption test [paste].
Give one sentence of evidence per line, mark any score without evidence as 0, and name the weakest line.
```

Any 0 on truthfulness and compliance stops the idea.

## Sample prompt: Platform test (Grok Bot, Claude, ChatGPT)

```
Assume [model vendor] added this feature as a default button tomorrow.
List what my buyer would still need from me. If the list is empty, say so plainly.
```

An empty list means timing, not a business.

## Free preview of workflow-not-wrapper-ai-product-offer-and-pricing-kit.md

````markdown
# Workflow, Not Wrapper: test an AI product idea, shape the offer and set the price

A prompt pack and working method by PaperChaseWebb for anyone about to build or sell something on top of a model. A seven-score absorption test, a one-page workflow product spec, a problem mining loop, a channel translation table, a model column check, an offer ladder with pricing rules, a claim check for the success stories you are copying, a fidelity gate for generated media, a failure table, a worked example, a rules block and 14 prompts. Plain .md. Works in Grok Bot, Claude, ChatGPT, Claude Code or Codex.

Most "AI business" advice stops at the model call: pick a tool, put a nice screen on it, charge for it. That screen is the easiest part of the business to copy, and the model vendor can absorb it in a single release. What lasts is the work around the call: the buyer's inputs, the template, the review step, the rules, the delivery format, the records and the channel that brings the buyer back. This pack is about building and pricing that part.

This is a decision method, not financial or legal advice, and not a promise that any idea will sell. Every revenue figure you hear about someone else's product is a claim until you check it.

## Contents

1. The wrapper problem on one page
2. The absorption test: seven scores
3. The workflow product spec
4. Problem mining: read and record, never seed
5. Channel translation: which launch tactics survive your buyer
6. The model column check
7. The offer ladder and pricing rules
8. Check the story before you copy the model
9. The fidelity gate for generated output
10. Failure modes and fixes
11. Worked example: listing videos for real estate agents
12. Rules block and 14 prompts
13. Sources

## 1. The wrapper problem on one page

A wrapper is a clean interface over one model call. It can make money for a while, usually because it got there first or found a channel others missed. It has three weak points:

- **The vendor can add it.** If the model provider makes your feature a default button, your customers have a reason to leave the same week.
- **A competitor can rebuild it in a weekend.** The prompt and the screen are visible from the outside.
- **You inherit someone else's roadmap.** Price changes, policy changes, outages and model retirements land on you without notice.

A workflow product uses the model as one replaceable part inside a job the buyer already does. The buyer pays for the finished result inside a process they recognize, not for access to a prompt box.

Quick test before anything else: **if the platform added your feature tomorrow, what would your buyer still need you for?** If the honest answer is "nothing", you have timing, not a business. Timing can still pay, but price it and plan your exit from it as timing.

## 2. The absorption test: seven scores

Score each line 0, 1 or 2. Write one sentence of evidence next to every score. A score with no evidence counts as 0.

| # | Score | 0 | 1 | 2 |
|---|---|---|---|---|
| 1 | Platform absorption | The vendor's default chat already does it | The vendor could add it as a button | It needs steps, inputs or rules the vendor will not build for one niche |
| 2 | Workflow depth | One input, one output | Several steps, but no review or delivery | Inputs, template, review, approval and delivery in the buyer's format |
| 3 | Data that improves results | None | Saved preferences | Buyer records or past outcomes that visibly make the next result better |
| 4 | Distribution access | You hope people find it | You can reach some buyers | You already sit where buyers gather, or they come back on their own |
| 5 | Model independence | Welded to one model or vendor feature | Swap possible with real rework | Model swap is a config change and a test run |
| 6 | Unit economics | Cost per job unknown | Known, thin margin | Known, with room for retries, refunds and support |
| 7 | Truthfulness and compliance | Output can mislead and nobody checks | A check exists but is optional | A check and a human approval are part of the product |

**Reading the total (0 to 14):**

- **0 to 5:** a wrapper. Build it only as a cheap experiment or a free tool that leads somewhere.
- **6 to 9:** a product with a hole. Name the lowest line and fix that first.
- **10 to 14:** worth a spec. Go to section 3.

Any 0 on line 7 stops the idea until it is fixed, whatever the total. A product that can mislead a buyer's customer is a liability with a checkout button.

## 3. The workflow product spec

One page. If you cannot fill it, you do not yet know what you are building.

```
# Workflow product spec: [name]

Buyer: [one job title in one kind of business]
The job today: [what they do now, in their words, and how long it takes]
Finished result: [what they hand to their own customer or boss]

Inputs: [files, forms, records the buyer already has]
Template or rules: [the format, brand rules, legal lines that never change]
Model step: [what the model does; name it as replaceable]
Review step: [who checks what, against what, before anything leaves]
Approval: [who says yes, and how that yes is recorded]
Delivery: [format, place, and whether it lands inside a tool they already use]
Records: [what is kept, for how long, and who can see it]

Moat lines (from the absorption test): [the two highest-scoring lines and why]
Weakest line: [the lowest score and the fix]
Not doing: [what this product refuses to do]
Cost per finished job: [model, storage, retries, support minutes]
```

Two rules for the spec:

- **The model step is one line.** If the model is most of the page, you are describing a wrapper.
- **"Not doing" is never empty.** It is where the product stops being "videos for anyone" and becomes something a specific buyer trusts.
````

## Worth knowing

- Scores are judgment with evidence, not math.
- Rules for outreach, disclosure, payments and generated media differ by market and change often. Check the current rules for yours.
- No method guarantees sales.

## Questions

**Is this only for software products?** No. It works for services and installs too. The ladder in section 7 runs from a free self-check to a monthly managed service.

**Does it tell me what to build?** No. Ideas come from your complaint log. The pack helps you test, shape and price them.

**How is this different from Agent Cost Levers?** Agent Cost Levers cuts what agent runs cost you. This pack decides whether an AI product is worth building and how to price it.

## How to buy

$29 one time at https://paperchasewebb.ai/bots/marketplace/items/workflow-not-wrapper-ai-product-offer-and-pricing-kit/. Stripe checkout, then an instant download page; the download link is also in your emailed receipt.

Seller: PaperChaseWebb Inc., mgmt@paperchasewebb.com