Rob Dull › AI Portfolio
AI Portfolio

PM tools built for
practice and demos

AI-native PM tools covering the full product workflow. Each works standalone — and each can hand its output to the next.

Network Diagram ↗ Explore AI ProdOps ↗
How the AI works: Five tools call the Claude API directly from your browser with your own key (you'll be prompted on first run; the key lives in session memory and clears when the tab closes). No key? Every tool still works — sample runs, worked examples, instant layouts, or the editor itself.
One connected workflow
01 — DISCOVER
Persona generator
Open ↗
02 — MAP
Journey map
Open ↗
03 — VET
Feature intake
Open ↗
04 — SPEC
Docs generator
Open ↗
05 — ALIGN
Hoshin + V2MOM
Open ↗
06 — FUND
Business case
Open ↗
07 — MEASURE
Ops radar
Open ↗
Each tool hands its output to the next
Enterprise User Research
JTBD + Enterprise
Persona Generator
Live

Takes a rough user description and produces a detailed enterprise persona with Jobs to Be Done statement, org context, tech comfort, key interactions, and relationship friction, detailed enough to anchor a journey map.

Where it fits in the workflow
Step 01 — Discover. The upstream anchor of the entire network. Every downstream tool produces better output when it has a grounded persona to work from. Used standalone: AI-augmented synthesis from interview notes. In a connected workflow, the Research ingest MCP pre-populates the tool with structured themes from Dovetail or Notion interview repositories.
AI-augmented
Jobs to Be Done Persona card Key interactions Relationship friction Info preference Copy to export
Open tool ↗
PERSONA CARD Marcus Chen Software Engineer · PetHealth Member "I just want to know it's working." Until he needs it. JOBS TO BE DONE When I want to So I can TECH COMFORT Mobile App Web Portal Phone / Chat
What it produces
  • Full persona card: name, title, department, tenure, reports-to, location
  • First-person quote, bio, wants and needs, frustrations
  • Technology comfort ratings across five relevant platforms
  • Information preference: how this person consumes data and why it matters for design
  • Key interactions table with color-coded friction, dependency, and reporting badges
  • Organizational impact of unmet job: three quantified consequences
  • Copy-to-clipboard export formatted for Miro, Figma, or product briefs
    • Customer Experience
    Customer
    Journey Map Builder
    Live

    A journey map is only useful if it reflects what actually happens, not what teams assume. This tool provides an editable, stage-by-stage canvas for mapping customer actions, goals, touchpoints, emotions, and organizational responses, with a draggable sentiment curve and one-click AI export for analysis.

    Where it fits in the workflow
    Step 01 — Discover. Produces the pain cluster map that feeds the opportunity merge agent. The JSON export is structured for direct consumption by downstream tools or an AI agent. Used after the persona generator, a journey map grounded in a JTBD persona is significantly more predictive than one built from assumptions. The signal aggregator MCP can enrich journey stages with live support ticket themes.
    AI-augmented
    Editable stages User sentiment curve Touchpoints Pain points Business goals Export for AI
    Open tool ↗
    PRE-ENROLL ACCOUNT MGMT CLAIMS RENEWAL ACTIVITIES TOUCHPOINTS SENTIMENT KPIs 📱 🌐 📋 ✉️
    What it produces
  • Draggable sentiment curve that captures emotional highs and lows across the journey
  • Row-level data for activities, goals, touchpoints, pain points, and business OKRs
  • Toggleable rows: show only the layers relevant to your current audience
  • Add and remove stages dynamically without losing other row data
  • One-click plain-text export structured for pasting directly into an AI prompt
  • JSON export for version control or integration into other tools
    • Strategy and Execution
    Hoshin Kanri
    + V2MOM Alignment
    Live

    Enter your V2MOM (vision, values, methods, and measures) and this tool cascades it directly into a Hoshin Kanri X-matrix. Values become annual objectives. Methods map exactly to activities. Measures drive key metrics. The matrix is color-coded by value group, correlation dots are interactive, and Claude can align and refine the whole thing on demand.

    Where it fits in the workflow
    Step 02 — Align. Provides the strategic alignment context that every other tool depends on — what the organization is trying to accomplish and why. Used standalone: the clearest way to cascade a V2MOM into an actionable X-matrix. In a connected workflow, the Strategy context MCP pre-populates the tool from your Confluence or Notion planning wiki, and the strategy cascade agent flags OKR alignment gaps after each run.
    AI-augmented
    V2MOM cascade X-matrix visualization Color-coded by value Interactive correlation dots AI alignment Vision alignment row
    Open tool ↗
    HOSHIN KANRI + V2MOM X-MATRIX VISION ALIGNMENT Redesign claim submission flow Proactive renewal outreach Launch in-app vet directory Member sentiment dashboard Priorities and Activities Annual Objectives Key Metrics Long-term Objectives A seamless, trustworthy member experience across every stage CO-OWNERS
    What it produces
    • A full Hoshin Kanri X-matrix cascaded directly from your V2MOM: values to objectives, methods to activities, measures to KPIs
    • Color coding by value group flows across activity rows, objective columns, and KPI columns for instant visual correlation
    • Vision alignment row showing how directly each annual objective expresses the stated vision
    • Interactive dot cells: click to cycle through direct, complementary, and empty correlation states
    • AI cleanup and realignment that improves phrasing, adjusts correlations, and preserves your manual edits
    • Adjustable matrix dimensions: add or remove rows and columns for any zone on the fly
    Business Justification
    Technical Business
    Case Generator
    Live

    Translates project inputs into a structured, stakeholder-ready business case in three analysis depths: quick pitch, standard, or full analysis with ROI modeling.

    Where it fits in the workflow
    Step 03 — Justify. Sits at the business case approval HITL gate between the opportunity map and the feature intake pipeline. Used standalone: AI-augmented — the PM drafts and edits. Connected to the signal aggregator MCP and strategy context MCP: the agent enriches financial claims with live benchmarks before the PM reviews.
    AI-augmented → AI-native
    Quick Pitch Standard Full Analysis ROI calculation Salary-based productivity Risks of inaction
    Open tool ↗
    TECHNICAL BUSINESS CASE FULL ANALYSIS EXECUTIVE PITCH 3.2x ROI 18 mo. payback PEOPLE IMPACT RISKS OF INACTION COST / BENEFIT
    What it produces
  • Cost / benefit analysis with salary-based productivity ROI and payback period
  • End-user and employee impact section tied to the people you enter
  • Risks of inaction: financial, reputational, and talent retention
  • Technical complexity and delivery confidence assessment
  • How returns and risks evolve over time, month-by-month narrative
    • Agentic Workflow Demo
    Feature Intake
    Pipeline
    Live

    Four sequential AI calls where each step's structured output becomes the next step's input. Takes a raw feature idea through intake triage, JTBD framing, assumption mapping, and a RICE-based prioritization brief. An explanation panel at each step shows the design decision behind the chain.

    Where it fits in the workflow
    Step 04 — Intake. The most AI-native tool in the set. It is itself an agentic chain, not just an AI-assisted form. Output feeds directly into backlog grooming and the three parallel review gates (architecture, security, business) before any engineering commitment. When connected to the Intake source MCP, approved briefs write back to Jira automatically.
    AI-native
    Agentic chaining JTBD framing Assumption mapping RICE scoring Human checkpoints
    Open tool ↗
    01 TRIAGE 02 JTBD 03 ASSUMPTIONS 04 BRIEF CORE PROBLEM SIGNAL STRONG JOB STATEMENT UNDERSERVED EMOTIONAL ASSUMPTIONS HIGH MED LOW RECOMMENDED ACTION SPIKE RICE SCORE: 42
    What it produces
    • Structured problem statement: core problem, who is affected, what they cannot do today
    • Three-layer JTBD analysis: functional, emotional, and social job dimensions
    • Assumption map ranked by uncertainty × impact, each with a fastest-test method
    • RICE scoring rationale with recommended next action: spike, prototype, backlog, or reject
    • Explanation panel at each step showing the design decision behind the chain
    Agentic PDLC Demo
    PDLC Docs
    Generator
    Live

    Four sequential AI calls that turn a feature brief into delivery-grade documentation: a classified requirements register (FR/NFR/BR/DR/CON with MoSCoW priorities and explicit/inferred/assumed confidence markers), a proposed solution outline with measurable NFR thresholds, and two finished documents — a developer spec and a plain-language user guide with structured screenshot placeholders — every section traced back to numbered requirement IDs.

    Where it fits in the workflow
    Design & requirements. The step between intake and delivery. One click on a finished Feature Intake brief carries it here; the tool applies full-scale PDLC disciplines — numbered requirement IDs, open questions instead of invention, NFR evidence plans, end-to-end traceability — compressed into a browser chain. On completion it saves a requirements register that a future backlog-decomposition tool will consume.
    AI-native
    Requirements register MoSCoW + confidence NFR thresholds Traceability index Screenshot placeholders .md download
    Open tool ↗
    01 REGISTER 02 OUTLINE 03 DEV DOC 04 USER GUIDE REQUIREMENTS REGISTER FR-001 M EXPL FR-004 M INFR NFR-002 M INFR NFR-003 S ASSM DR-001 S ASSM OPEN QUESTIONS OQ-001 BLOCKING OQ-002 Developer Documentation API ENDPOINTS POST /api/exports Realizes: FR-001, NFR-005 User Guide Step 1: Open the export drawer SCREENSHOT PLACEHOLDER · SHOT-01 Capture: drawer open over the data view
    What it produces
    • Requirements register: FR/NFR/BR/DR/CON with IDs, MoSCoW, and confidence markers
    • Open questions and a conflict scan instead of invented certainty
    • Proposed solution outline: file tree, API surface, NFR thresholds and evidence plans
    • Developer spec and user guide as downloadable markdown, cross-referenced both ways
    • Deterministic traceability index mapping every requirement to where it is realized
    Organizational Health
    Product Operations
    Health Radar
    In development

    A structured self-assessment across the core dimensions of product operations delivery. Surfaces where teams are strong, where they are struggling, and what to prioritize first.

    Where it fits in the workflow
    Step 05 — Measure. The feedback layer that closes the loop. After delivery, it surfaces whether the ops system is healthy enough to sustain the next cycle. Used standalone: AI-augmented self-assessment with prioritized recommendations. In a connected workflow, the Metrics MCP auto-populates the delivery and tooling dimensions from live DORA metrics and sprint data, so the PM fills in only the qualitative dimensions.
    AI-augmented → AI-native
    Delivery health Radar visualization Prioritized recommendations Team assessment
    PRODUCT OPS HEALTH RADAR IN DEVELOPMENT Discovery Delivery Metrics Tooling Collab Strategy TOP PRIORITIES Discovery Delivery Metrics Tooling Collab Strategy Focus: 2 areas flagged
    What it will produce
    • Radar chart across eight product operations dimensions: discovery, delivery, metrics, tooling, and more
    • Scored self-assessment with dimension-level commentary
    • AI-generated interpretation of the pattern: what the shape of the radar means organizationally
    • Prioritized recommendation set ranked by impact and effort
    • Export for leadership review or team retrospective facilitation
    Background reading · AI-augmented → AI-native
    These tools exist on a continuum, from AI-augmented to AI-native product operations.
    Each tool here can stand alone as an AI-augmented workflow: faster synthesis, better structured output. But they were designed to connect: persona informs journey map, journey map feeds opportunity mapping, business case gates the intake pipeline, intake brief drives strategy alignment. At full connection, with agents and MCP threading the workflow, that's AI-native product ops. This piece maps where each tool sits, how the architecture hangs together, and what it means for the PM role.
    Explore AI ProdOps ↗
    Side Projects
    Built for fun, engineered with care.

    Not everything here is a PM tool. This is what I build when the constraint is a friend group and a weekend, not a product roadmap.

    Browser Game
    500 Card Night
    Live

    500 is the card game I've played with the same group of friends since college. It's simpler than bridge, more interesting than euchre. The existing online versions are buggy, clumsily monetized, and don't follow our house rules. I've wanted to build my own since I was working in the Java games division at Nokia. That was twenty years ago. I finally built it in two days with Claude.

    No accounts, no server, no database. Share a room code; play from anywhere. Four players, peer-to-peer, entirely in the browser.

    Try it solo (no friends required)
    Open the game in four separate browser tabs. Each tab joins the same room and plays as a different seat. It's how I test every change: the full four-player game running on one machine, no network required.
    TypeScript React / Vite WebRTC / PeerJS Three-repo architecture No backend
    Play the game ↗
    500 Card Night lobby
    500 Card Night play view 500 Card Night end of hand 500 Card Night bidding 500 Card Night bidding round 2