Product Specs, Roadmaps, and Research Decisions
Turn customer evidence into testable PRDs, dependency-aware roadmaps, and stakeholder updates with clear decisions and tradeoffs.
Before & after
Same prompt, same agent — the only difference is the skill.
A PRD is a great way to clarify requirements before building. Writing a good one involves understanding the problem you're solving, who needs it, and how you'll measure success. To create a PRD that the engineering team will actually follow, I'd need to understand your product better. What's the core use case for this chatbot? Who are the users, and what problems does it solve? Are there competitive chatbots out there doing something similar? How will this fit into your existing product? I can help you structure a PRD with sections for requirements, success metrics, and user stories, but the specific content really depends on your business goals and constraints. Maybe we could start with a problem statement?
PRD: AI-Powered Customer Support Chatbot
OBJECTIVE: Reduce support queue time by 40% and improve CSAT from 78% to 85% in 90 days
PROBLEM STATEMENT:
Customer support receives 300+ queries daily; 60% are repetitive questions (password reset, billing, how-to). Current response time: 4-8 hours. Support team burned out. Solution: Chatbot handles Tier-1 questions, escalates complex issues to humans.
USER STORIES:
US-1: Customer asks FAQ question
AS A customer
I WANT to get an immediate answer
SO THAT I don't wait for support team
ACCEPTANCE CRITERIA:
→ Chatbot responds within 2 seconds
→ Answer matches knowledge base (100% accuracy required)
→ User can escalate to human with one click
→ Interaction logged for training
US-2: Support agent reviews escalated tickets
AS A support agent
I WANT to see chatbot conversation history
SO THAT I don't repeat information
ACCEPTANCE CRITERIA:
→ Full chat transcript visible in ticket view
→ Questions already asked by chatbot highlighted
→ Escalation reason ("I don't understand") shown
US-3: Manager monitors chatbot performance
AS A support manager
I WANT a dashboard showing resolution rate by category
SO THAT I can identify gaps in knowledge base
ACCEPTANCE CRITERIA:
→ Daily/weekly resolution rate (goal: 65%)
→ False answer rate tracked (goal: <2%)
→ CSAT for chatbot-handled cases measured separately
→ Alert if escalation rate exceeds 40%
SCOPE:
IN (90-day MVP):
→ Password reset, billing, how-to categories (350 questions)
→ Text-only (no voice, no video)
→ Integration with Zendesk (escalations auto-create tickets)
→ Admin dashboard for knowledge base updates
OUT (Future phase):
→ Sentiment analysis
→ Multi-language support
→ Voice/video support
→ Predictive escalation
SUCCESS METRICS:
→ Resolution rate: 65% (measured: "did customer close ticket without agent response")
→ CSAT improvement: 78% → 85%
→ Response time: <2 seconds
→ False answer rate: <2%
→ Adoption: 60% of new customer inquiries hit chatbot
→ Support time saved: 20 hours/week
RISKS & MITIGATION:
→ Risk: Chatbot gives wrong answer → Mitigation: 2-week human review of all responses before launch
→ Risk: Support team resists (fear of layoffs) → Mitigation: Redeploy team to higher-value work (custom issues)
→ Risk: Low adoption → Mitigation: Add positive feedback loop ("thanks for rating this helpful")
TIMELINE:
Week 1-2: Collect FAQ data, train model, build admin UI
Week 3-4: Human review, Zendesk integration, team training
Week 5-6: Phased rollout (10% of traffic, monitor)
Week 7-12: Full launch, monitoring, iteration
DEPENDENCIES:
→ Zendesk API access (dev team)
→ QA for edge cases (support team, 20 hours)
→ Marketing for launch comms (1 week notice)
OWNER: Product Manager (Alice)
ENG LEAD: Backend engineer (Bob)
APPROVED: CTO, VP SupportAbout this skill
Your PRD lists features but leaves engineering guessing. Your roadmap says everything is next, and customer feedback never becomes a clear decision. This skill connects the evidence, requirements, sequencing, and stakeholder message so each deliverable explains what to do and why.
What it does
- Writes PRDs with bounded scope, prioritized requirements, observable acceptance criteria, and defined success metrics.
- Builds Now/Next/Later, quarterly, or OKR-aligned roadmaps with dependencies, capacity constraints, and explicit commitment levels.
- Synthesizes interviews, surveys, and support tickets into traceable themes without treating repeated tickets as independent customers.
- Drafts executive, engineering, sales, and customer updates around impact, decisions, risks, and asks.
- Reviews product metrics for denominator errors, cohort maturity, and unsupported causal claims before recommending action.
- Compares competitors by customer job and workflow, separating verified capabilities from claims and unknowns.
What you get
A finished draft sized to your request: a PRD, roadmap, research synthesis, competitive brief, metrics readout, or stakeholder update. Material assumptions, evidence gaps, and the next decision are explicit, with links between artifacts when useful.
Best for
PMs and product leads turning scattered customer evidence into buildable scope, defensible priorities, and credible updates for cross-functional teams.
SKILL.md preview
---
name: product-specs-roadmaps-and-research-decisions
description: "Create and revise PRDs, roadmaps, research syntheses, competitive briefs, metrics reviews, and stakeholder updates by tracing evidence to decisions, testable scope, dependencies, and outcomes. Use when the user says \"write a PRD for this\", \"our roadmap has too many priorities\", \"turn these interviews into opportunities\", \"engineering keeps asking what this requirement means\", \"draft a product status update\", \"why did activation drop\", or \"compare these competitors\"."
---
# Product Specs, Roadmaps, and Research Decisions
… (sign up to view the full skill)More business & ops skills
View all Business & Ops skills →Product Practice Diagnostic
Trace a product-team symptom (unused features, roadmap churn, priority fights) to the broken stage of the loop and fix one thing first.
Feature Launch Go-to-Market Plan
Plan one feature launch: tier, plan access, audience messages, readiness gates for support, sales, billing, and docs, and success metrics.
Pricing Tier and Packaging Diagnosis
Finds why your pricing tiers underperform and redesigns value metric, fences, price gaps, and upgrade triggers with a migration plan.
Audience Panel Review for Drafts and Ideas
Simulates 3-5 real people from your audience to react to a draft before launch, with tagged evidence, ranked edits, and follow-up questions.