Support

Decision Gates

Turn your validated decision patterns into explicit rules that guide your AI's behavior.

Available on Standard tier and above.

What Are Decision Gates?

Decision Gates are explicit rules derived from your validated Instincts. While Instincts are learned patterns that influence AI behavior subtly, Decision Gates are promoted rules that actively guide how your AI operates.

Think of the difference between a habit (Instinct) and a policy (Decision Gate). Both shape behavior, but a policy is deliberate and documented.

How Gates Emerge

Decision Gates don't appear from nowhere. They follow a path:

  1. Observation — Tempreon notices a behavioral pattern in your sessions
  2. Instinct formation — The pattern is validated across multiple observations
  3. Promotion — When an Instinct reaches high confidence, it can be promoted to a Decision Gate
  4. Active enforcement — The Gate actively guides AI behavior in relevant situations

The "Promoted from Instinct" indicator on each Gate links back to the pattern it came from.

What a Gate Contains

Each Decision Gate includes:

  • Name — A short label for the rule
  • Description — What the rule does, in plain language
  • Evidence Summary — The observations that support this rule
  • Confidence — How strongly the evidence supports this pattern
  • Sample Size — How many signals contributed
  • Domain — Whether the rule applies globally or to a specific Domain
  • Status — Active, Draft, Paused, or Retired

Viewing Your Gates

The Decision Gates page shows:

  • Active Gates — Currently influencing your AI's behavior
  • Inactive Gates — Draft, Paused, or Retired gates for reference

Each Gate displays its confidence level and the evidence behind it.

Try Saying...

  • "What decision rules are active for my consulting domain?" — See Domain-specific gates
  • "Create a rule: always include ROI projections in client proposals" — Define a new Gate
  • "What guidelines are guiding how you respond to me?" — Surface active Gates

Use Case: Codifying Best Practices

Rachel is a COO who noticed her AI kept suggesting aggressive timelines. She had told it multiple times that she preferred conservative estimates with buffer. After enough corrections, Tempreon promoted this into a Decision Gate: "When estimating timelines, add 20% buffer and flag any aggressive assumptions." Now, every timeline recommendation from her AI automatically includes the buffer — she doesn't need to correct it anymore. The pattern became policy.