RISEN
The procedural framework for tasks where the steps matter as much as the answer.
RISEN extends the original RISE framework (Role, Input, Steps, Expectation) with an explicit 'Nuances' slot to capture edge cases. It became standard at consulting firms because client-facing AI workflows have to be auditable — every step has to be inspectable, every edge case has to be documented, and the deliverable shape can't drift.
Every RISEN prompt has these sections.
Role
The persona / expertise lens — 'a forensic accountant', 'a staff SRE', 'a DTC growth marketer with 8 years at unicorns'.
Input
The raw material the model operates on — the document, dataset, transcript, code file.
Steps
The explicit ordered procedure. Numbered. Each step a verifiable action, not vague advice.
Expectation
The exact deliverable shape — three columns, 5 bullets, 80–120 words, no preamble.
Nuances
The edge cases, exceptions, gotchas, and tie-breaker rules the model must respect.
- →Multi-step analysis pipelines (data → cleanup → analysis → recommendation).
- →Code generation with required architectural conventions.
- →Document processing where the procedure must be followed in order.
- →Any task where 'how the model got there' matters for audit or compliance.
- ×Single-shot generative tasks where structure would just slow it down.
- ×Tone-heavy creative writing (use CO-STAR).
- ×Exploratory brainstorming where you don't yet know the steps.
From one sentence to a RISEN section.
"Analyze our last 30 days of NPS comments and tell us what to fix."
## Steps 1. Cluster comments into themes using verbatim phrases. 2. For each cluster, count detractors vs passives vs promoters. 3. Rank clusters by (detractor_count × avg_severity). 4. For the top 5 clusters, propose one concrete product or process fix…
Ready to write your first RISEN prompt?
Drop a sentence. The Forge analyzes your task, asks the right clarifying questions, then generates a complete enterprise meta prompt structured exactly to RISEN.