Back to portfolio
Prompt Engineering · Product

ODI Innovation Advisor

Outcome-Driven Innovation Advisor

Teams struggle to define real customer needs → this AI enforces Jobs-to-be-Done thinking and outcome statements → users get structured, measurable innovation inputs instead of vague ideas.

01The Problem

Most innovation efforts fail because teams don’t have a clear, shared definition of customer needs. They rely on ideas, opinions, or loosely defined “pain points,” which leads to misaligned product decisions and unpredictable results. Without a consistent way to define and prioritize needs, organizations end up guessing rather than systematically identifying what will create value.

02What the AI Does

I guide users through the Outcome-Driven Innovation (ODI) process by structuring thinking around jobs-to-be-done and desired outcomes. I: * Define and refine jobs-to-be-done using strict formatting rules * Generate properly structured desired outcome statements * Translate user inputs into measurable needs (time / likelihood metrics) * Guide segmentation thinking based on unmet outcomes * Apply the opportunity algorithm conceptually to prioritize needs * Reframe vague ideas into ODI-compliant inputs I am built on a GPT-5.3 language model with custom instructions that enforce ODI principles, including strict syntax for outcome statements and a process-oriented approach to innovation. I also reference embedded ODI knowledge frameworks derived from foundational materials like and the JTBD canvas .

03Design Decisions

01 · Choice

Constrained the AI to Tony Ulwick’s ODI methodology

Why

To replace vague, general-purpose outputs with a rigorous innovation framework grounded in measurable customer needs

Constraint

Prevents drifting into generic brainstorming or idea generation without defined outcomes

02 · Choice

Enforced strict outcome statement syntax (time / likelihood formats only)

Why

ODI requires standardized, quantifiable need statements to enable prioritization and segmentation

Constraint

Eliminates ambiguous or non-measurable “needs”

03 · Choice

Prioritized jobs-to-be-done over personas, demographics, or features

Why

ODI defines markets around jobs, not customer types or solutions

Constraint

Blocks common but flawed approaches like persona-driven innovation

04 · Choice

Structured responses around the ODI process steps

Why

To ensure users move from definition → outcomes → segmentation → prioritization systematically

Constraint

Reduces flexibility for freeform exploration in favor of process discipline

05 · Choice

Explicitly prohibited hallucination and speculation about outcomes or usage

Why

To maintain credibility and ensure outputs are grounded in user-provided inputs and ODI logic

Constraint

Limits ability to “fill in gaps” without sufficient information

06 · Choice

Embedded coaching behavior (guiding, reframing, correcting)

Why

ODI requires precise thinking; users often start with poorly framed inputs

Constraint

May challenge or reject user phrasing rather than accommodating it

07 · Choice

Avoided tool integration beyond language reasoning

Why

[Creator: add rationale]

Constraint

Cannot perform real quantitative analysis, surveys, or statistical segmentation

05Key Insight

AI becomes far more valuable when constrained by a rigorous thinking framework than when left as a general-purpose generator.