Roadmap Narrative Skill

Turn a roadmap — however rough — into polished, audience-specific narratives. One input, three outputs.

Input: $ARGUMENTS

Accepts:

• /roadmap-narrative — Will ask for the roadmap interactively
• /roadmap-narrative [path/to/roadmap.md] — Read from file
• /roadmap-narrative [pasted content] — Accept pasted text, table, or bullet list

If no input is provided, ask:

"Paste your roadmap or describe what you're shipping over the next 1-3 quarters. It can be rough — bullet points, a table, or even just a list of feature names. I'll do the rest."

If input is brief (just feature names), ask ONE follow-up question:

"What's the strategic theme tying this roadmap together? What problem are you fundamentally solving for customers this period?"

Don't ask more than one question. Work with what you have.

Phase 1: Parse the Roadmap

Extract from the input:

• Items: Each initiative, feature, or theme
• Time horizon: Now/Next/Later, Q1/Q2/Q3, or whatever structure is given
• Implied theme: What's the common thread across this work?
• Notable gaps: Anything conspicuously absent that stakeholders might ask about?

If no time horizon is given, group items into Now / Next / Later based on apparent priority or sequence.

Phase 2: Identify the Narrative Arc

Every roadmap has a story. Find it:

• Problem arc: "Customers couldn't do X. We're fixing that."
• Platform arc: "We're laying foundations now so we can move fast later."
• Growth arc: "We're doubling down on what's working."
• Catch-up arc: "We're closing the gap on table-stakes functionality."
• Strategic pivot: "We're moving from X to Y."

Name the arc explicitly — it becomes the spine of every version of the narrative.

Phase 3: Generate Three Narratives

Generate all three versions. Each should be self-contained — do not require the reader to have read another version.

Version 1: Executive Summary (100-150 words)

Audience: CPO, CEO, board. Assumes they haven't read anything about this.

Structure:

1. The strategic theme — What's the one thing this roadmap is about?
2. The customer problem — Why now? What changed?
3. The three big bets — What are we actually shipping? (3 items max)
4. The outcome — What will be measurably different in 6 months?

Format: Prose, not bullets. Sounds like something you'd say in a meeting, not a status update.

Version 2: Engineering / Team Brief (300-400 words)

Audience: Engineering leads, designers, PMs. They want context and sequencing.

Structure:

1. Why this roadmap — The strategic rationale and key decisions made
2. Sequencing logic — Why are we doing things in this order?
3. Dependencies & risks — What are we betting on? What could derail us?
4. How we'll know it's working — Success metrics per initiative (use Amplitude data if available)
5. What we're NOT doing — Explicit out-of-scope to prevent scope creep

Format: Mix of prose and bullets. Opinionated. Name trade-offs.

Version 3: Customer-Facing / External (150-200 words)

Audience: Customers, sales, CSMs. They want to know what's coming that helps them.

Structure:

1. The customer problem we're solving — Empathetic, outcome-focused
2. What's coming — Framed as customer outcomes, not feature names
3. Timing — Honest but not committed. Use "coming soon", "this quarter", "later this year"
4. Call to action — What should customers do with this info? (Try the beta, talk to their CSM, etc.)

Format: Friendly, non-technical. No internal jargon. Sounds like a product newsletter, not a Jira epic.

Phase 4: Output

Present all three versions back-to-back, clearly labelled.

After presenting, offer:

"Want me to adjust the tone, emphasis, or length on any of these? I can also draft a Confluence page, Slack announcement, or slide deck outline if useful."

Tips

• If the roadmap is heavily feature-named ("Add bulk export", "New dashboard widget"), translate feature names into customer outcomes before writing any version.
• If you spot a strategic gap or incoherence in the roadmap (e.g., two conflicting themes, or a quarter with nothing customer-facing), flag it gently: "One thing I noticed — the Q2 work is all infrastructure with no customer-facing output. Do you want me to flag that in the exec version or omit it?"
• Never invent initiatives that weren't in the input. If the roadmap is thin, say so — don't pad it.