Thought Leadership

Thought Leadership Manager

Manage the full lifecycle of vendor-neutral thought leadership pitches: generate, track, draft, and close.

Commands

Parse the first argument to determine the command:

If no argument is provided, prompt the user to choose a command.

Shared: Pitch memory

All commands read from pitch-output/.pitch-memory.json. If it doesn't exist, initialize with:

{
  "pitches": [],
  "last_run": null
}

Each entry in pitches has: id (sequential int), date, topic, angle, outlets, output_file, status ("open" | "writing" | "accepted"), draft_file (null or path).

Command: pitches

Research trending developer topics in the AI space that overlap with Postman's product areas, then generate 5 vendor-neutral pitches for contributed content at tech publications.

Voice and framing rules

1. Vendor neutral. Pitches focus on the trending topic, industry problem, or developer challenge. Do not mention Postman, any competitor, or any specific product by name. The pitch sells an expert perspective on a topic, not a product.
2. One paragraph per pitch. Each pitch is a single paragraph (4-6 sentences) that frames the trend, states the angle, and hints at the insight the article would deliver. It should make an editor want to read the full piece.
3. Expert voice. Write as a senior technologist proposing a byline. Authoritative, opinionated, specific. No marketing language ("unlock", "supercharge", "revolutionize", "leverage").
4. Actionable angles. Each pitch should propose a clear, differentiated angle. "AI agents are changing APIs" is too broad. "Why your API error messages are the biggest blocker to agent adoption" is specific enough.

Workflow

1. Load pitch memory

Read pitch-output/.pitch-memory.json.

2. Load Postman AI topic areas

Read references/postman-ai-topics.md to understand which AI topic areas overlap with Postman's capabilities. These are the topic guardrails, not content to include in pitches.

3. Research trending topics

Run parallel WebSearch queries to find what's trending in AI and APIs right now.

If a focus area was provided, tailor all queries toward that focus. Otherwise use these defaults:

Batch 1 (run in parallel):

• "AI agents API" trending 2026 — agent-to-API interaction trends
• "agentic AI" enterprise challenges 2026 — enterprise adoption friction
• "API security AI agents" 2026 — security concerns with AI calling APIs
• "MCP protocol" OR "model context protocol" developer — MCP ecosystem growth
• "AI gateway" OR "agent gateway" API management — infrastructure for AI traffic

Batch 2 (run in parallel):

• "API observability AI" OR "agent observability" — monitoring AI-driven API calls
• "AI SDK" OR "AI developer tools" trends — tooling for AI app builders
• "API governance AI" compliance enterprise — governance in the age of agents
• site:techcrunch.com OR site:thenewstack.io AI agents API 2026 — what outlets are already covering

After searches complete, use WebFetch on the top 2-3 most relevant results to get deeper context on the freshest trends.

4. Cross-reference and select topics

Compare the trending topics against the Postman AI topic areas from Step 2. Select 5 topics where:

• The topic is genuinely trending (multiple recent sources, not just one article)
• It overlaps with at least one Postman AI capability area (so the company could credibly author it)
• The angle hasn't been pitched before (check against pitch memory from Step 1)
• Same topic with a different angle is allowed and not considered a duplicate

5. Generate pitches

For each of the 5 selected topics, write:

1. Topic: 3-5 word label (e.g., "Agent-to-API authentication gaps")
2. Pitch: One paragraph (4-6 sentences). Frame the trend, state the angle, hint at the insight. Vendor neutral. No product names.
3. Suggested outlets: 2-3 publications that cover this space and accept contributed content. Read references/outlets.md for the outlet reference.
4. Why these outlets: One sentence per outlet explaining why it fits.
5. Postman overlap: One sentence (internal note, not part of the pitch) noting which Postman AI area connects. This helps the PR team position the author.

6. Write output

Save to pitch-output/pitches-YYMMDD.md using this format:

# Thought Leadership Pitches

**Generated:** [date]
**Pitches:** 5 | **New topics:** [count] | **New angles on existing topics:** [count]

---

## 1. [Topic label]

**Pitch:**
[Single paragraph pitch text]

**Suggested outlets:**
- [Outlet 1] — [why it fits]
- [Outlet 2] — [why it fits]
- [Outlet 3] — [why it fits]

**Internal note:** Overlaps with [Postman AI area]. Potential author: [suggest role, e.g., "CTO", "Head of AI", "Developer Advocate"].

---

[Repeat for pitches 2-5]

7. Update pitch memory

Append each pitch to pitch-output/.pitch-memory.json with a sequential id, status: "open", and draft_file: null.

Update last_run to the current ISO 8601 timestamp. Prune entries older than 6 months (except those with status "writing" or "accepted").

8. Report results

Print a summary:

Pitches written to pitch-output/pitches-YYMMDD.md

Topics covered:
  1. [Topic] → [Outlet 1], [Outlet 2]
  2. [Topic] → [Outlet 1], [Outlet 2], [Outlet 3]
  ...

Previous pitches on file: [count]

Command: list

Display all pitches from pitch-output/.pitch-memory.json in a table:

ID | Date       | Topic                              | Status   | Outlets                    | Draft
---|------------|-------------------------------------|----------|----------------------------|------
1  | 2026-08-14 | Agent-to-API authentication gaps    | open     | TechCrunch, The New Stack  | —
2  | 2026-08-14 | API error messages block agents     | writing  | InfoQ, DZone               | pitch-output/draft-2.md
3  | 2026-07-20 | MCP server testing strategies       | accepted | The New Stack              | pitch-output/draft-3.md

Group by status: Open first, then Writing, then Accepted. Show counts per status at the bottom.

Command: write

Write a full contributed article draft from a selected pitch.

Workflow

1. Read pitch-output/.pitch-memory.json and show only pitches with status "open" or "writing" in a numbered list.
2. Ask the user to select a pitch by number.
3. Read the pitch's output file to get the full pitch details (topic, angle, outlets, Postman overlap).
4. Read references/outlets.md to understand the target audience for the suggested outlets.
5. Research the topic further with 3-4 targeted WebSearch queries to gather supporting data, examples, and recent developments.
6. Write a 1200-1800 word contributed article draft that:
   • Matches the pitch angle exactly
   • Is vendor neutral (no product names, no Postman)
   • Uses an expert byline voice — opinionated, backed by specifics
   • Opens with a hook that frames the problem or trend
   • Includes 2-3 concrete examples, data points, or scenarios
   • Ends with a forward-looking take, not a sales pitch
   • Uses subheadings to break up the piece
   • Targets the audience of the first suggested outlet
7. Save the draft to pitch-output/draft-[id]-[slugified-topic].md with frontmatter:

---
pitch_id: [id]
topic: [topic]
target_outlet: [primary outlet]
word_count: [count]
status: draft
---

# [Article title]

[Article body]

8. Update the pitch in .pitch-memory.json: set status: "writing" and draft_file to the path.
9. Print a summary with the file path and word count.

Command: accepted

Mark a pitch as accepted and complete.

Workflow

1. Read pitch-output/.pitch-memory.json and show pitches with status "open" or "writing" in a numbered list.
2. Ask the user to select a pitch by number.
3. Update the pitch in .pitch-memory.json: set status: "accepted".
4. Print confirmation with the topic and outlet.

Done when

The requested command completes successfully and pitch memory is up to date.