HOW I BUILD WITH AI — WORKFLOW TEMPLATES

Ashley Kays / The Wealth Experiment

Reusable teaching patterns; no implied earnings or measured time savings.


SHARED BRIEF
Audience:
Problem:
Output:
Source links and dates:
Scope / exclusions:
Deadline:
Cash and total-hour limits:
Acceptance criteria:
Approvals:


HANDOFF
Owner:
Inputs:
Output location:
Checks completed:
Uncertainty or blockers:
Next owner and next action:


COST LOG
Shared subscription allocation:
Variable tool charges:
Hosting/assets:
Setup, failed attempts and review hours:
Accepted outputs:
What is still unmeasured:



BUILD AND PUBLISH A WEBSITE WITH AGENTS

Real example: the AI income experiment guide on this site was authored, connected to the training and blog pages, checked, published through GitHub and Railway, and verified on the live site. The steps below turn that pattern into a reusable website workflow. They do not claim a measured time saving or commercial result.

1. Brief the job
Owner: Human + product lead
Action: Write the audience, visitor problem, one primary action, required content, exclusions and definition of done. For a service page, define what a visitor must understand before contacting you.
Output: A one-page brief with acceptance criteria.
Review: Can someone describe the purpose without seeing a mockup?
My notes: ______

2. Inspect what already exists
Owner: Research + developer
Action: Read the current site and source before changing anything. Map the existing navigation, forms, content sources and deployment path. Record what to reuse.
Output: A current-state note and a bounded change list.
Review: Are you editing the source that generates the page, rather than an output that the next build overwrites?
My notes: ______

3. Design one journey
Owner: Designer + content worker
Action: Sketch the path from headline to evidence to action. Use Figma if a visual prototype will resolve a question. Prepare realistic content and an image brief. Keep text out of images when it needs to remain selectable.
Output: A page outline or prototype and reviewed copy.
Review: Can users find the offer, understand the next step and use it on a small screen?
My notes: ______

4. Build a reviewable version
Owner: Development worker
Action: Give the builder the approved brief and existing design rules. Ask for the smallest change set, meaningful error states and accessible controls. Keep credentials out of public code.
Output: A preview or branch with an explanation of the changes.
Review: Do links, forms, keyboard focus and empty states work? Does the page preserve existing behavior?
My notes: ______

5. Review and publish
Owner: QA + human approver
Action: Run relevant build and link checks. Inspect the actual page, including mobile layout. Confirm the intended audience and deployment target before publishing. Keep a known prior version for rollback.
Output: A reviewed release and live URL.
Review: Does the live URL contain the intended version and serve its images and downloads? A successful build alone is not live verification.
My notes: ______

6. Observe and improve
Owner: Measurement worker + human
Action: Choose one meaningful action such as a completed inquiry. Record visits and qualified actions, with consent-aware analytics where configured. Ask users where they hesitate.
Output: An observation log and one next improvement.
Review: Do not call a page successful because it is live. Missing evidence remains unknown.
My notes: ______

STARTER PROMPT
Use this approved website brief and the existing project files. First identify the content source and current user journey. Propose the smallest implementation. Return changed files, checks run, unresolved issues and publication steps. Do not invent testimonials, results or unavailable features. Ask before actions outside the approved scope.

ASSIGNMENT
Create one offer page or improve one existing page. Stop when the intended visitor can understand the offer and complete the primary action. Do not expand into a complete application to finish this exercise.


CREATE A USEFUL COURSE WITH AN AGENT TEAM

Teaching example: build a short “Run your first AI income experiment” course around the guide already published here. The six-session outline is available; this walkthrough explains how to develop and pilot a fuller course. It is not a claim that this expanded course has been delivered or that learner results have been measured.

1. Define the learner and outcome
Owner: Human instructor + research worker
Action: Name the learner’s starting point, prerequisite skills and final task. For example: complete an experiment brief with a buyer, offer, budget, evidence threshold and next action. Interview potential learners about where they get stuck.
Output: A learner brief and observable learning outcome.
Review: Is the outcome something learners can demonstrate, rather than “understand AI”?
My notes: ______

2. Design backward from the task
Owner: Curriculum worker
Action: Write a rubric before the lessons: specific buyer, evidence versus assumptions, bounded test, cost limit and decision rule. Break the final task into teachable steps.
Output: A lesson sequence and assessment rubric.
Review: Does every lesson help the learner complete the task? Remove tool tours that do not.
My notes: ______

3. Research and storyboard
Owner: Research + content worker
Action: Use primary documentation for tool instructions and keep a source/date log. For each lesson outline the problem, demonstration, guided practice, independent task and review.
Output: A storyboard with verified claims and example inputs.
Review: Can another instructor follow it? Are hypothetical numbers and examples labeled?
My notes: ______

4. Make the learning assets
Owner: Designer + content worker
Action: Create readable screenshots from real demonstrations, original diagrams, a workbook and a completed example. Add text descriptions for visuals and transcripts for recorded instruction. Use Figma when visual layout needs refinement.
Output: A lesson, worksheet, answer example and accessible visual.
Review: Does the image show the actual step? Is the text readable on a phone and in print?
My notes: ______

5. Pilot with learners
Owner: Human instructor + QA worker
Action: Invite suitable testers, explain the pilot scope and observe whether they can complete the task without rescue. Track unclear instructions, time spent and errors. Ask permission before collecting identifiable feedback or recordings.
Output: Pilot notes, rubric scores and a revision list.
Review: Completion is not mastery. Can the learner repeat the task with a different example?
My notes: ______

6. Publish and maintain
Owner: Publishing worker + human
Action: Check navigation, downloads and any enrollment steps. Explain prerequisites, contents and access terms. Keep an update log for changing interfaces and sources.
Output: A working lesson path and maintenance checklist.
Review: Do not promise certification, income or included subscriptions unless those are actually provided.
My notes: ______

STARTER PROMPT
Design a practical lesson for this learner and outcome. Return prerequisites, an observable objective, a five-part storyboard, guided practice, an independent task and a rubric. Use my sources for product facts, mark gaps and label invented examples. Do not claim that learners achieved results. Keep the first lesson small enough to pilot.

ASSIGNMENT
Build one complete lesson, not ten outlines. Include a worked example, a blank worksheet, a learner task and a rubric. Use learner evidence to decide whether the next lesson is ready to make.


RESEARCH AND TEST AN INCOME IDEA WITH AGENTS

Illustrative experiment: offer a bounded proposal-preparation pilot for independent consultants. This is a teaching scenario, not a new service listing or an earnings claim. The existing income guide supplies a hypothetical 14-day test and a cost example.

1. Find a reachable buyer
Owner: Human + research worker
Action: List skills you can deliver, audiences you can reach and recurring problems you have observed. Read relevant public sources and conduct voluntary customer conversations.
Output: A shortlist with evidence, sources and access routes.
Review: Is the problem based on behavior, or only an AI-generated persona?
My notes: ______

2. Choose the risky assumption
Owner: Product worker
Action: Separate whether the buyer cares, will pay, can be reached and can be served sustainably. Choose the uncertainty that could invalidate the idea first.
Output: One hypothesis and the evidence that would disprove it.
Review: Do you need an interview, prototype, paid pilot or delivery test? Do not build software to answer a buying question.
My notes: ______

3. Make a small sample
Owner: Designer + delivery worker
Action: Create one realistic sample with fictional or authorized data. Describe the deliverable, exclusions, price hypothesis and acceptance criteria.
Output: A sample and a plain-language offer.
Review: Can you deliver the promised outcome, including review and corrections, within the effort limit?
My notes: ______

4. Run one distribution test
Owner: Marketing worker + human
Action: Choose one appropriate channel. Draft an invitation or useful demonstration with a clear next step. Personally review publication and outreach. Record the audience and source.
Output: Approved drafts and a record of relevant invitations or visits.
Review: Respect declines and channel rules. Likes are not purchases; absent traffic does not disprove demand.
My notes: ______

5. Deliver and count the costs
Owner: Delivery + QA worker
Action: If a buyer agrees, fulfill the defined pilot and obtain acceptance. Record all setup, selling, delivery and review time, plus direct costs and refunds.
Output: Accepted work and a complete evidence ledger.
Review: Include failed attempts. Do not label revenue as profit or ignore your time.
My notes: ______

6. Continue, change or stop
Owner: Analyst + human decision-maker
Action: Compare observations with your preselected decision rule. Explain missing data and alternative explanations. Select the next smallest test, or stop at the agreed limit.
Output: A decision memo and next review date.
Review: A few purchases are directional evidence, not a stable conversion rate or proof of repeatable wealth.
My notes: ______

STARTER PROMPT
Evaluate this income idea using only my skills, customer notes and sources. Separate facts, assumptions and unknowns. Identify the riskiest assumption and propose one small test with a cash limit, total-hour limit, primary metric, denominator and continue/change/stop criteria. Do not invent demand, market size, customer quotes or expected earnings.

ASSIGNMENT
Complete one experiment card and one sample. Set the limits before reaching buyers. Record the outcome honestly, including a decision to pause.


FINAL REVIEW
What exists:
What was checked:
What remains unknown:
Continue / change / stop:
Next smallest step:


https://wealthexperiment.live/how-i-build-with-ai/