From Will (Xiaohongshu UX/AI job search blogger)'s course Let's Start section — the entry point and roadmap for the full material set. This page only distills content exclusive to the beginner navigation: the mindset and usage rules before reading the materials, the capability landscape of the full set, and the order beginners should follow.How to use the bot, interview question banks, and prompts for turning projects into stories are already fully covered in Will's AI bot usage mindsets + general UX interview questions + creative project explanation prompts, this page does not rewrite them, and only points to that resource at relevant places.
Significance for PMs: How the author defines this material set and what mindset he asks you to use it with is itself a model of product onboarding. It repeatedly emphasizes one stance —AI provides possibilities, humans make trade-offs; unjudged output is merely "seemingly complete noise". This statement also holds true when you use AI to write PRDs / do competitive analysis / write retrospectives: a tool being able to run through a process does not mean its output is usable; value comes from your filtering, correction, and restructuring.
1. 6 must-read usage tips before viewing the materials (most easily skipped, read this first)
Why this is listed separately: These 6 rules are the "instruction manual" the author provides before the main content. Beginners most often dive straight into a stage tutorial, only to end up using the wrong mindset (treating it as an "AI efficiency automation tool") or wasting time on the wrong module. Align with these 6 stances first to avoid detours later.
| # | Rule | One-sentence explanation (including Why) |
|---|---|---|
| 0 | This material set only serves UX career advancement (promotion/job change) | The goal is the content quality of your portfolio + verbal expression (interviews/presentations), not studying flashy AI workflows or pursuing efficiency automation. Before using it, confirm that what you want is "to nail your portfolio and interviews", not "to collect AI gimmicks". |
| 1 | Generate AI images in English first, then switch to Chinese | Whenever you use AI to generate images (UI, charts), add "output in pure English" at the end of the prompt first, then replace the text with Chinese after getting the result.Why: The process of switching from the English version to the Chinese version forces you to spot issues with visual weight and information hierarchy (English typography exposes whether the hierarchy is clear). Text content is not subject to this rule. |
| 2 | Prompts in the documents must be "downloaded locally" before use, do not open them directly | Click the six-dot icon on the left of the Prompt → download to local → copy for use. There are two types of Prompts:direct copy-paste type, and requires you to fill in information type (for the latter, follow the hints in parentheses, e.g. "custom project background required"). |
| 3 | Which parts to read in each of the four modules— this rule ensures you do not waste time | There are four columns under each stage tutorial, with functions in order: ①Process Process & Framework (can be skipped) ②Cognition Cognition —the direction in which you direct AI; this part must be mastered thoroughly③ AI Skill Technique Hands-on steps for AI implementation ④References Visual result reference (what the finished product looks like, for you to manually make into your own work).Why: Cognition is the part where "you make judgments", and is the core of this method; Process is actually the least important. |
| 4 | If your Prompt writing skill is still at the level of "can you make it more systematic / can the structure be more detailed" | don't read anything else, go learn Google Prompting Essentials Specialization(Google's official prompt course) first. This is an explicit prerequisite set by the author. |
| 5 | When two tools appear together in tutorials (e.g. GPT and Claude) = choose one, prioritize Claude | The optimal choice is usually Claude. When you encounter places where "either GPT or Claude works", default to Claude. |
2. Capability landscape of the full material set (5 pillars + how to combine them)
Why: The most common confusion for beginners is "which of these many modules should I use". The author splits the full set of capabilities into 5 pillars, then provides3 ready-made combination recipes. Look at the landscape first, then pick a combination based on your goal (building a portfolio / quickly validating an experience); you do not need to learn everything.
5 core capability pillars
| Pillar | What it is | Corresponding material section / next step |
|---|---|---|
| ➊ AI builds portfolio framework and narrative | Use AI to produce portfolio structure + copy + verbal talking points | For usage of the bot (PitchPro), see Will's AI bot usage mindsets + general UX interview questions + creative project explanation prompts; Section "3" of this page covers an even earlier resume-tailoring skill |
| ➋ AI generates detailed content | User research support processes + visualized data + interface and interaction solutions, covering AI techniques for each of the 6 UX stages | For an overview of the 6 stages, see Overview of Will's AI-assisted UX workflow (full-stage skeleton from A-Discover to F-Deliver + navigation to 12 sister wiki pages)(Discover & Research → Define & Insight → Ideate → Experience Design → Validate → Deliver) |
| ➌ Vibe Coding pipeline | End-to-end 0→1 Vibe Coding (using natural language to drive production of runnable products) workflow, accessible even without a development background | VibeCoding / Figma Make section of the materials |
| ➍ Specialized AI capability system | Build the fundamentals of AI capabilities (Vibe Coding + Prompt writing + systematic AI tool courses), embed AI into real UX practice to form a closed capability loop | Google Prompting / Tools section of the materials |
| ➎ Agent + Skill + MCP work system | Use Agent (Codex) to read online/local files, and enable interconversion between PRD ↔ code ↔ interface ↔ UI system in Figma. The author calls this the currently most efficient UX+AI system | Codex + Figma section of the materials |
Terminology notes: Vibe Coding = "chatting out" runnable code in natural language without writing code manually; Agent = an AI that can read files, call tools, and execute multi-step tasks on its own; Skill = a packaged, reusable piece of AI capability (like a plugin); MCP = a standard interface that lets AI connect to external tools/data (can be understood as "the USB port for AI").
3 ready-made combination recipes
| Recipe | Applicable scenario | Output |
|---|---|---|
| ➊+➋+➎ Full closed loop | Build a "showable + tellable" portfolio from 0 to 1 | Complete portfolio (structure+content+visuals) / high-quality copy and talking points for each project / every page includes data charts, user research evidence, UI interaction solutions / visuals benchmarked against industry excellence / every project has a demo to present. For complete cases, see the materials F - Deliver → Case Study |
| ➋+➌ Quick experience validation | Quickly build a "complete product experience" in daily UX work | Understand value conversion between business ↔ UX / define requirement points and pain points / apply AI at every UX node / AI generates UI + interactions / Vibe Coding converts to a real product / launchable, testable, quickly validatable |
How to choose: If you are building a job search portfolio → take the full closed loop (➊+➋+➎); if you only want to quickly test an experience solution in daily work → take quick validation (➋+➌). You do not need to learn everything at the start.
3. An even earlier step before PitchPro: resume-tailoring skill (one resume per job)
Why this is listed separately:Will's AI bot usage mindsets + general UX interview questions + creative project explanation prompts covers using PitchPro to write portfolio copy and interview talking points; but before that, the author also provides a resume tailoring skill, which is an independent step not covered in the sister page. The logical order is:First use it to revise your resume to match the target job posting → then enter PitchPro to build your portfolio and talking points.
resume-tailoring-skill-master.zip is a skill (packaged reusable capability) that runs in the Codex or Claude Code client, usage:
- Download the zip → click the plus sign in Codex/Claude Code to load it → run it once.
- Two ways to run it:
- Method 1 (works for both Codex / Claude Code): Trigger with a single prompt —
使用 resume-tailoring skill,帮我针对这个岗位定制简历:[粘贴岗位 JD];我的简历库在:[上传附件]. - Method 2 (recommended for Claude Code): Paste directly
[岗位 JD] + [上传简历] + "走一遍完整流程", and the Agent will ask follow-up questions step by step, helping you see what is missing from your resume.
- Method 1 (works for both Codex / Claude Code): Trigger with a single prompt —
- What it does: Rewrite your resume according to the target job JD, analyze JD match rate, research target company culture,dig out experiences you did not include in your resume, output Word/PDF, and build a reusable "resume bank".
Transferability for PMs: This is the paradigm of "using an Agent's follow-up questioning to complete information" — instead of letting AI output everything in one go, let it ask you item by item for the missing details. This is the exact same approach as letting AI act as a "questioner" when you clarify requirements and write PRDs.
4. Recommended beginner route (stringing the above into an onboarding order)
- First read the 6 rules in Section "1" of this page → align your mindset (especially: this serves portfolios/interviews, the Cognition module must be learned, choose Claude between the two tools).
- Make up for gaps in Prompt skill first → Google Prompting course (Rule 4).
- Pick a combination recipe based on your goal(Section "2" of this page): take ➊+➋+➎ for job search portfolios; take ➋+➌ for daily experience validation.
- Resume first(Section "3" of this page): use the resume-tailoring skill to revise your resume to match the job posting.
- Enter the main portfolio and interview process → full bot mindsets, interview question banks, creative explanation prompts are at →Will's AI bot usage mindsets + general UX interview questions + creative project explanation prompts.
- To see the distribution of AI techniques across the 6 UX stages → Overview of Will's AI-assisted UX workflow (full-stage skeleton from A-Discover to F-Deliver + navigation to 12 sister wiki pages).
Key terminology quick reference (Glossary)
| Term | One-sentence explanation |
|---|---|
| Let's Start section | The entry/roadmap layer of the full material set, the object distilled on this page |
| Cognition module | The column among the four modules that covers "the direction in which you direct AI", the core the author explicitly says must be learned |
| English first then Chinese for image generation | The fixed routine for AI image generation: generate in pure English first, then switch to Chinese, use the switch to expose visual hierarchy issues |
| 5 core capability pillars | Portfolio narrative / detailed content / Vibe Coding / AI capability system / Agent+Skill+MCP |
| resume-tailoring skill | Resume-tailoring skill running on Codex/Claude Code, one resume per job, the prerequisite step before PitchPro |
| PitchPro | Custom GPT built by Will, for writing portfolio copy and interview talking points — for usage see Will's AI bot usage mindsets + general UX interview questions + creative project explanation prompts |
<!-- Ingest note: This page overlaps with [Will's AI bot usage mindsets + common UX interview questions + creative project explanation prompts](/knowledge/kb-6adc9eed3d2ec7e5f869c762f213811b) by approximately 30% (both cover the PitchPro bot mechanism for writing portfolios and interview content); all overlapping content points to the companion page and is not rewritten here. Content unique to this page = 6 usage tips + 5-pillar capability landscape + 3 combination recipes + resume-tailoring skill + beginner route. Image gaps: The Let's Start section contains about 7 capability pillar schematic diagrams, and the bot subdirectory contains about 30 operation screenshots; all are interface/schematic images, the core methods have been textualized, and images are not embedded (already added to.gitignore). Video gap: Let's Start includes a 3:33 video with audio track (March 29_(1).mp4, 66MB) demonstrating the practical effect of PitchPro producing portfolio frameworks and interview talking points; this agent cannot read the audio track so it has not been transcribed; if it is an introductory walkthrough, supplementary transcription can be considered later. -->