Defining Dimensions and Facilitation Guides
The core building blocks of a Signal Event — what they are, how they work, and how to review them
Every Signal Event is built from two core components: dimensions (what you're measuring) and facilitation guides (how you're exploring each dimension). Understanding what each one does — and what makes each one good — helps you design better events and evaluate AI-generated designs effectively.
Dimensions
A dimension is a specific, measurable aspect of the topic you're researching. It's the defined thing you're trying to learn about through the conversation.
A good dimension is: - Specific enough to be measurable — ""competitive positioning"" is too broad; ""how clearly the rep can articulate their product's advantage over the primary competitor in this territory"" is a dimension - Important enough to justify the conversation time — if you can't describe what a high score versus a low score looks like, the dimension may not be ready - Distinct from other dimensions — dimensions that substantially overlap produce redundant data
Each dimension is scored on a 1–5 scale after the session. A score of 1–2 is Low, 3 is Moderate, 4–5 is High. When there isn't enough evidence to score reliably, the system returns an abstention (""No Score"") rather than guessing.
Once your event is complete, the Research Provenance panel shows the evidence basis for each AI-generated dimension: whether it's grounded in research literature, derived from domain knowledge, or flagged as having thin evidence. Use this to evaluate how confident you should be in each dimension after the event has been generated.
Facilitation Guides
A facilitation guide is the conversation plan for a dimension. It tells Savo how to explore the dimension — what to probe for, what evidence counts as signal, and when the topic is sufficiently covered.
Each guide includes: - Goal — what the guide is designed to learn - Prompt Seeds — the opening approaches and follow-up directions Savo uses to explore the topic - Saturation Criteria — the specific evidence signals that indicate the dimension has been adequately explored - Threshold Overrides — how much evidence is needed before moving on (sufficient / abbreviated / immediate thresholds) - Time Overrides — minimum and maximum time parameters for this guide
When reviewing AI-generated guides, ask: does this approach actually surface the evidence needed to score this dimension? Would a skilled human interviewer probe this way? Do the saturation criteria capture what ""good coverage"" actually looks like?
Reviewing an AI-Generated Event
When Event Studio generates a design, evaluate it against these questions:
- Does each dimension map clearly to a research question you care about?
- Does the facilitation guide for each dimension actually produce the evidence needed to score it?
- Are the saturation criteria realistic — specific enough to be meaningful, not so demanding they’re never met?
- Does the Research Provenance panel show strong grounding for the most important dimensions?
If the answer to any of these is uncertain, edit directly or regenerate with targeted feedback before publishing.