# AI Faculty Atlas User Manual

This guide is for applicants. It focuses on **how to use the product to make better application decisions**, not on technical details.

---

## 1) What this product helps you do

AI Faculty Atlas is a workflow for one core question:

**Which professors and labs should I spend my application time on?**

You can use it to:

- Explore AI faculty across schools
- Identify likely fit and obvious mismatches quickly
- Build a shortlist of professors
- Compare shortlisted professors side by side
- Open full professor dossiers before outreach
- Draft outreach messages after evidence review
- Run a readiness plan and close profile gaps

The product is decision support. It does not guarantee admissions outcomes.

---

## 2) Fast start (first 15 minutes)

If you only do one pass, do this:

1. Open `ai-atlas.html`
2. Set your background in **Your background (powers relevance sort)**
3. Keep sort on **Relevance to me**
4. Apply basic chips:
   - `Recruiting`
   - `Hide do-not-rank`
   - `Evidence prep` (optional)
5. Open 6–10 promising profiles in dossier view
6. Pin 3–4 professors to **Shortlist**
7. Go to `shortlist.html` and run **Compare**
8. End with `readiness.html` to get a concrete next-step plan

---

## 3) Main pages and when to use each one

## Explore (`ai-atlas.html`)

Use this when you are still searching broadly.

- **Search bar**: professor, school, topic keywords
- **School filter**: narrow to one program
- **Sort**:
  - `Relevance to me`: best default after setting your background
  - `Decision evidence / curated quality`: evidence-first browsing
- **Primary chips**:
  - `All`
  - `Shortlist`
  - `Recruiting`
  - `Hide do-not-rank`
- **Advanced filters**: placement, signal quality, startup/thesis proxies

Use Explore to reduce hundreds of options to a manageable set.

## Dossier (`professor-dossier.html?id=...`)

Use this when deciding if a specific professor deserves outreach effort.

Read in this order:

1. Decision card summary (top)
2. Placement and evidence sections
3. Research fit notes and benchmarks
4. Mentoring/recruiting signals
5. Outreach draft area (if you are ready)

Rule of thumb: if evidence remains thin after dossier review, do not over-invest in that target.

## Shortlist (`shortlist.html`)

Use this to move from intuition to comparison.

- Keep 2–4 active targets for side-by-side compare
- Track tradeoffs explicitly (fit, recruiting, placement profile)
- Remove anyone who looks weak under direct comparison

This page is where final targeting quality usually improves most.

## Readiness (`readiness.html`)

Use this before writing many applications.

- Input your stage and assets
- Get prioritized gaps (what to improve first)
- Follow sequence, not everything at once

Treat readiness as execution planning, not as a score.

## Resources (`applicant-decision.html`)

Use for policy/process context:

- Wellbeing navigation
- PhD playbook references
- Culture/visa guidance

This page supports decisions; it does not replace school-specific official policies.

## Lab decision (`lab-decision.html`)

Use only if you are evaluating stay/switch/leave decisions inside a current program.

---

## 4) How to read signals correctly

### Recruiting status

- `Recruiting (public)` means there is public evidence of open recruiting
- `Not recruiting (public)` means public evidence says closed/not taking
- `Recruiting unknown` means no strong public confirmation either way

Unknown is not automatically bad. It means you need better verification.

### Data quality tiers

- **Evidence-backed / curated**: stronger confidence
- **Auto-generated / limited signal**: use with caution
- **Do-not-rank**: avoid using as a primary decision anchor

Prefer fewer high-quality targets over many low-quality ones.

### Decision card copy

Decision cards summarize public evidence in plain language.
Use them as a first-pass frame, then verify with dossier details.

---

## 5) Suggested workflow by stage

### Stage A — Early explorer

- Goal: understand landscape
- Spend most time in Explore
- Build a loose shortlist of 8–12

### Stage B — Focused applicant

- Goal: improve target quality
- Cut shortlist to 4–6 using dossier + compare
- Prioritize recruiting-open and higher-evidence profiles

### Stage C — Outreach execution

- Goal: high-quality, customized contact
- Final list of 3–4 priority targets
- Draft outreach only after dossier evidence review

---

## 6) Common mistakes to avoid

- Skipping background setup, then trusting relevance sort
- Applying to too many low-signal targets
- Treating one metric (for example placement prestige) as the whole decision
- Drafting outreach before reading full dossier evidence
- Confusing unknown recruiting status with explicit rejection

---

## 7) A practical weekly routine

Use this 60–90 minute weekly loop:

1. Refresh Explore filters and sort
2. Add/remove shortlist candidates
3. Deep-read 2 dossiers
4. Compare top candidates
5. Update readiness plan and one concrete action

You should end each week with one better decision, not more tabs.

---

## 8) Sign-in and access notes

- Sign-in uses magic link (email-based)
- If purchasing is temporarily unavailable, continue with free browsing/decision workflow
- During free phases, you should not need purchase flow to complete core exploration and compare steps

---

## 9) What “good usage” looks like

After one solid session, you should have:

- A profile-informed shortlist
- 2–4 compared priority professors
- A clear next action for your application readiness

If you only remember one principle:

**Use evidence to narrow choices, then spend effort where uncertainty is lowest and fit is strongest.**