Skip to content

AI & Product Innovation

Building better decisions with AI.

I explore how Generative AI can be embedded into product discovery, requirements engineering and decision-making — not simply used to generate content.

Explore IREM + C3P0 See my product approach

AI should not replace product thinking. It should amplify it.

The opportunity is not simply to automate the production of artifacts. It is to improve the quality, consistency and traceability of the decisions that create those artifacts.

A personal framework

Where AI creates value in Product Management

My own perspective on how AI adds value to product work — not an industry framework.

  1. Level 01

    Augment

    Use AI to accelerate individual product activities.

    Examples

    • Research synthesis
    • Documentation
    • Analysis
    • Ideation
    • Communication
  2. Level 02

    Structure

    Use AI to improve the consistency and quality of product decisions.

    Examples

    • Problem framing
    • Requirements analysis
    • Acceptance criteria
    • KPI definition
    • Dependency analysis
    • Prioritization support
  3. Level 03

    Transform

    Design product workflows where AI becomes part of the operating model.

    Examples

    • AI-enabled product discovery
    • Decision-support systems
    • AI-assisted requirements engineering
    • AI-native workflows

Case in practice

IREM + C3P0

From requirements engineering to AI-enabled decision support.

IREM combines Agile Product Management, Requirements Engineering and structured decision models with Generative AI. C3P0 operationalizes this methodology inside the product workflow.

  1. Product thinking
  2. IREM methodology
  3. Structured decisions
  4. C3P0
  5. AI-enabled workflow
  6. Measurable outcomes
Conceptual relationship — not a technical architecture.

What C3P0 does

AI as structured decision support

C3P0 supports and guides the team; product decisions remain with people. Integrated into Atlassian Jira through Rovo AI.

  • Requirements

    Development-ready User Stories

  • Validation

    BDD acceptance criteria

  • Testing

    Business-oriented test scenarios

  • Measurement

    Business KPI definition

  • Dependencies

    Dependency analysis

  • Readiness

    Definition of Ready validation

  • Decomposition

    Epic decomposition and feature slicing

  • Traceability

    Value traceability

What makes this different

Not just AI-generated content.

Generative AI as a content tool
  1. Generate text
  2. Review output
  3. Edit
  4. Use
AI as decision support
  1. Understand context
  2. Challenge assumptions
  3. Structure decisions
  4. Validate completeness
  5. Generate artifacts
  6. Measure outcomes

A conceptual contrast between two ways of using Generative AI — not a description of how AI is usually used.

My product approach

How I think about AI in Product

  1. Start with the problem

    AI should solve a meaningful product problem.

  2. Structure before automating

    Automating a poorly defined process only makes the problem faster.

  3. Keep humans in the loop

    AI can support decisions, but accountability remains with people.

  4. Measure outcomes

    AI initiatives should have measurable product or business outcomes.

  5. Design the workflow

    The biggest opportunity often comes from redesigning the process around AI rather than adding AI to an existing step.

AI and Product Discovery

AI changes the shape of Product Discovery.

Generative AI can reduce the cost of exploring alternatives, synthesizing information and challenging assumptions. That creates an opportunity for Product Managers to spend more time on problem framing, trade-offs and decision quality.

  1. Research
  2. Synthesis
  3. Hypotheses
  4. Alternatives
  5. Decisions
  6. Experiments
Conceptual flow.

AI and Requirements Engineering

Requirements are becoming an AI interface.

As AI becomes part of software delivery, the quality of the instructions, context, constraints and expected outcomes becomes increasingly important.

IREM explores this intersection by treating requirements not only as documentation, but as structured decision context that can be interpreted and operationalized by AI.

The perspective explored by the IREM initiative — not an established definition of requirements engineering.

Measured experiment

AI needs evidence.

João developed the IREM methodology and the C3P0 AI-enabled decision support approach. What was evaluated is the methodology — not AI alone.

+18%

Backlog items delivered

Teams using IREM delivered 18% more backlog items than teams using conventional requirements engineering.

100%

IREM-generated User Stories with explicit business success metrics

Compared with 76% in the traditional approach.

  • 8Agile teams
  • 4IREM + C3P0
  • 4Traditional approach
  • 3Months · controlled evaluation
Read the full IREM case

AI product maturity

From AI feature to AI-enabled product.

My conceptual model — not an external framework.

  1. AI Feature

    A single AI-powered capability.

  2. AI Capability

    Reusable intelligence embedded in a product.

  3. AI Workflow

    AI integrated into a repeatable end-to-end process.

  4. AI Product

    AI becomes part of the core value proposition.

  5. AI Operating Model

    AI changes how product and business decisions are made.

Toolkit

Tools are secondary. The problem comes first.

Technologies I have worked with, grouped by the job they do. Not all of them are AI tools.

Generative AI
Rovo AI (C3P0)
Product Discovery
IREMAtlassian Jira
Data & Analytics
SQLPower BI
Automation
APIsMicroservices
Engineering Collaboration
Atlassian JiraAzureAWS

Continuous learning

Supporting learning in AI and product practice.

  • IBM AI Product Manager
  • Product AI
  • AI MVP in 3 days
  • Design Thinking with AI

Explore the work

See AI applied to a real product problem.

Explore the IREM + C3P0 case study, including the problem, methodology, experiment and measured results.

Read the IREM case Back to selected work