03: Turning Pain Points into AI Challenges

Goal: Frame 1–2 high-quality, testable AI Challenge Statements from your top pain points.

3.1 Why This Step Matters

It’s easy to say “AI can help with engagement.” But without a clear, testable statement, the idea will collapse in later stages.

This section ensures you frame the right problem in the right way. It:

Keeps the beneficiary’s pain point at the center
Prevents vague or overly broad challenges
Forces you to check root causes, data, and constraints
Produces structured outputs that funders, partners, or AI developers can actually use

Every AI Challenge Statement must include five elements:

Element	Definition	Example
Problem Statement	Plain language description of the pain point	“Youth drop out of the rehab program after two weeks of attendance.”
Root Causes	Underlying factors contributing to the problem	“Notifications are generic, transportation barriers, low motivation.”
Desired Outcome	What positive result do we want?	“Increase retention to 80% at the 4-week mark.”
Constraints & Requirements	Limits, resources, or boundaries that must be respected	“Must protect patient privacy; must work on low-bandwidth mobile devices.”
Data Availability	What data do we have (or can get)?	“Attendance logs, app usage metrics, weekly counselor notes.”

Before finalizing a challenge, run it through the SMART+R filter:

Criterion	Question to Ask	✓ / ✗	Your Notes
Specific	Is the challenge clearly defined and focused on one core issue?
Measurable	Can we measure success with clear indicators or KPIs?
Achievable	Is the solution realistic given current resources and capacity?
Relevant	Does this align with the organization’s mission and SDG 3?
Time-bound	Can progress be seen within ≤12 months?
Responsible	Is the AI solution ethical, inclusive, & privacy-compliant?

❌ Weak Example: