Three Pillars

Optra Prism organizes its capabilities into three pillars, each serving a different audience.

Prompt Advisor

Audience: individual developers

Helps you write better prompts and improve your AI collaboration skills.

Feature	Description
Real-time PQ scoring	Scores every prompt before submission
Rewrite suggestions	Rewrites weak prompts with coaching tips
Anti-pattern detection	Flags retry storms, bundling, vague prompts
PRISM profiling	5-dimension score with trend tracking
Coaching notes	Dimension-specific improvement tips
Session replay	Review past sessions with per-prompt scoring

Usage Intelligence

Audience: developers and team leads

Provides visibility into how AI coding tools are used across sessions and teams.

Feature	Description
Token & cost tracking	Per-session, per-model, per-developer breakdown
Waste detection	7 patterns: retry storms, context bloat, model overkill, etc.
Throttle detection	Identifies rate-limiting events and calculates hours lost
Model rightsizing	Recommends cheaper models for simple tasks
Trend analytics	Usage patterns over days, weeks, months
Session explorer	Drill into any session to see turn-by-turn telemetry

Governance

Audience: team leads and platform engineers

Controls and policies for AI coding tool usage at the team and organization level.

Feature	Description
Budget caps	Soft warnings and hard blocks at spend limits
Guardrails	DLP, PII detection, prompt injection guard, content filter
Model access control	Restrict which models developers can use
Developer tiers	Assign spending tiers based on role or project
Audit logging	Full request/response capture for compliance
Recommendations	Data-driven policy suggestions with estimated savings

How they work together

Developer writes prompt
    → Prompt Advisor scores it, suggests improvements
    → Request routes through gateway (if enabled)
    → Governance applies budget + guardrails
    → Telemetry captured to ingest service
    → Usage Intelligence analyzes patterns
    → Recommendations generated
    → Developer improves → cycle repeats

The three pillars feed into each other: better prompts (Advisor) reduce waste (Intelligence), and data-driven policies (Governance) prevent budget overruns.