# Feature Request

**Status:** <mark style="color:$primary;">Live</mark>

## What It Does

Employees trigger the skill by mentioning a feature request, pain point, or broken tool in Claude. Claude conducts a 15-question conversational interview, synthesizes the answers into a branded scope document, calculates a Total Value Score, generates a downloadable DOCX, and submits an Asana task to the Director of Product once the employee saves the document to Google Drive and pastes the link back.

The employee describes the problem. Claude does the writing, the math, and the Asana submission. The Director of Product picks it up from there.

Examples of what the skill produces:

* A branded scope document (DOCX) with the Overview section fully populated from the interview.
* A Total Value Score calculated from the employee's Why Now and Importance selections.
* An Asana task in the correct platform project, pre-filled with custom fields and assigned to the Director of Product.

***

## Workflow

```
Employee describes a feature request or pain point → Department selection → 15-question conversational interview → Claude recommends Why Now + Importance → Total Value Score calculated → Draft presented in chat for review → Employee confirms → Branded DOCX generated → Employee saves to Google Drive → Employee pastes Drive link → Asana task created in platform's project → Confirmation
```

The skill uses a deterministic build script to generate the DOCX, ensuring every scope document matches the brand template exactly. Claude handles the conversation, the scoring, and the Asana submission. The employee handles saving to Drive and pasting the link.

***

## Interview Flow

The interview is 15 questions asked one at a time. Claude adapts naturally, follows up on vague answers, and skips ahead if the employee volunteers information early.

| Zone                  | Purpose                                                   | Questions                                                                           |
| --------------------- | --------------------------------------------------------- | ----------------------------------------------------------------------------------- |
| Platform & Context    | Identify which system is affected and what the request is | Platform selection, description of the problem or need                              |
| Users & Impact        | Understand who is affected and how broadly                | Who it's for, frequency and scale, current workarounds, impact of inaction          |
| Urgency & Success     | Gauge timing and define what good looks like              | Why now, what success looks like                                                    |
| Operational Details   | Capture access, notifications, reporting, and data needs  | Who should have access, notification requirements, reporting needs, data to capture |
| Integration & Rollout | Understand external dependencies and rollout approach     | External system touchpoints, phased vs. full rollout                                |
| Supporting Materials  | Collect screenshots, links, SOPs, or examples             | Employee can upload screenshots directly or paste links                             |

Claude follows strict interviewing principles: one question at a time, follow-up once on vague answers, redirect solution-prescribing back to problem description, and never force answers on questions that don't apply.

***

## Scoring Engine

After the interview, Claude recommends a **Why Now** and **Importance** selection based on what the employee described. The employee confirms or overrides.

**Why Now options:**

* A client asked for it
* It's blocking something else
* We have a workaround but it's painful
* It's a brand new capability
* Nice to have

**Importance options:**

* New Client Growth
* Expand Customer Spend
* Reduce Manual Work
* Reduce Security Risk

These feed into the Total Value Score formula:

**Total Value Score = (Potential Revenue x 0.6) + (Time Saved x 0.4)**

Potential Revenue maps from Importance (New Client Growth = 90, Expand Customer Spend = 75, Reduce Manual Work = 40, Reduce Security Risk = 30). Time Saved maps from Why Now (A client asked for it = 80, It's blocking something else = 70, We have a workaround but it's painful = 50, It's a brand new capability = 40, Nice to have = 10).

**Priority bands:**

* 70 to 100 = High
* 40 to 69 = Medium
* 0 to 39 = Low

The employee cannot change the Total Value Score directly. If they want to adjust the score, they change the Why Now or Importance selections and the score recalculates.

***

## Document Generation

Claude assembles a JSON object from the interview answers and passes it to a deterministic build script that generates the branded DOCX. The script owns all formatting: Navy title bar, Indicia Pink section headers, Proxima Nova fonts, Graphite body text, and consistent spacing throughout.

The generated DOCX contains:

* **Overview section** (employee-facing): Title, Status, Primary Requestor, Platform, Why Now, Importance, Total Value Score, Current State, As a Result (with impact bullets), Summary (who it's for, who it impacts, why it matters), and Supporting Materials.
* **Dev-owned sections** (placeholders for Product to fill): Solution, Resources Required, Technical Notes, Entry Point, Flow, Acceptance Criteria, In Scope, Out of Scope, LOE, Rock Association.

The employee only sees and reviews the Overview section. Dev-owned sections are present in the DOCX but are not shown during the chat draft review.

***

## Asana Routing

Each platform routes to its own Asana project. When the employee pastes their Google Drive link, Claude creates the task in the correct project with pre-filled custom fields.

| Platform  | Asana Project       | Section                | Key Custom Fields                                                      |
| --------- | ------------------- | ---------------------- | ---------------------------------------------------------------------- |
| Joii      | JOII                | Requirements Gathering | System, Task Type, Tag, Dev Status, Issue Priority, Submitted By       |
| Portal 2  | P2 Feature Requests | Submitted              | System, Task Type, Tag, Why Now, Importance, Value Score, Submitted By |
| Portal 1  | Portal 1            | Requirements Gathering | Task Type, Tag, Issue Priority, Submitted By                           |
| Automator | Automator           | Requirements Gathering | System, Task Type, Tag, Issue Priority, Submitted By                   |
| AI Tool   | AI Tools            | Requirements Gathering | Task Type, Tag, Issue Priority, Submitted By                           |
| JMP       | No Asana project    | N/A                    | N/A                                                                    |

**Portal 1 and AI Tool** do not set a System field because the project itself is the system.

**Portal 2** is the only platform that populates the Why Now, Importance, and Value Score fields on the Asana task.

**Joii, Portal 1, Automator, and AI Tool** derive Issue Priority from the Total Value Score band (High/Medium/Low).

**JMP** is being deprecated in favor of Joii. JMP submissions do not create an Asana task. Claude generates the DOCX and instructs the employee to send it directly to the Director of Product.

All tasks are assigned to Shannon Bannen (Director of Product). The Submitted By field is set to the authenticated Claude user after task creation.

***

## Edge Cases

**Vague requests:** Claude pushes back and asks for specifics. One or two rounds of clarification max. If the employee can't articulate the problem, they're asked to come back when they can name a specific friction point.

**Multiple unrelated requests:** Each request needs its own scope doc. Claude handles the most urgent one and asks the employee to come back for the rest.

**Solution-prescribing:** Claude redirects to problem description. The Overview is about what and why. Solution is the Director of Product's job.

**Skip the interview:** If an employee already has their thinking done, Claude accepts what they paste and structures it into the template, filling gaps with targeted follow-ups.

**Upset or venting employees:** Claude lets them vent briefly, then redirects into documenting the problem properly.

***

## Skill Components

| Component          | File                         | Description                                                                                                                       |
| ------------------ | ---------------------------- | --------------------------------------------------------------------------------------------------------------------------------- |
| Skill Instructions | SKILL.md                     | Complete conversation logic, interview questions, scoring formula, draft review flow, Asana routing rules, and edge case handling |
| Input Schema       | references/input-schema.md   | Exact JSON structure the build script expects, with field types and allowed values                                                |
| Brand Guide        | references/brand-guide.md    | Mailworks brand colors, fonts, and document styling rules                                                                         |
| Build Script       | scripts/build\_scope\_doc.js | Deterministic DOCX generator that applies brand formatting to the interview JSON                                                  |


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://knowledge.themailworks.com/skills/feature-request.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.