How it works
Six scenarios. Six dimensions. One report card.
SPARK does not test knowledge or use multiple choice. It tests how you reason in writing through situations that have no single correct answer, and scores each response against a published rubric.
The flow
Purchase access
One-time payment unlocks the assessment. You receive an access code by email after payment, usually within an hour. Current pricing: $97 (founding).
Complete six tasks
One open-ended scenario per capacity dimension. You write a response of any length; about 200-400 words is typical. Plain text only, no formatting required.
Receive your report
Scored per dimension on a 1-5 rubric with a behavior-level label. Strengths, gaps, and next steps for each dimension. Downloadable as PDF.
What scoring actually does
Each response is sent to an advanced AI algorithm running with temperature zero (deterministic) and a published rubric for the dimension being measured. The algorithm reads your response, applies the rubric anchors, and outputs a score from 1 to 5 plus:
- A behavior-level label that describes what reasoning at your score level typically looks like
- A multi-paragraph rationale explaining what specifically in your response led to that score
- Verbatim evidence quotes pulled from your own words
- A short list of strengths, gaps, and concrete next steps
Every scoring decision records the exact rubric version and model version that produced it. The decision can be re-evaluated if either the rubric or the model changes.
Time required
Roughly 30 minutes end to end. The system saves your progress between tasks, so you can pause and return as long as you keep the URL with your session id.
One thing worth knowing. A strong response shows your reasoning — name what is uncertain, generate alternatives, weigh them, commit to an action with a decision rule, and step back to consider what the situation reveals about the broader environment. The rubric rewards structure of reasoning, not length.