Open Questions — PRISM

Product / Strategy

1. Who is the first rollout user: AE, SE, BDR, partner manager, sales leader, or internal RevOps?
2. What is the minimum acceptable output for v1: full audit package, AE-only pre-call prep, or copilot research workspace?
3. Should PRISM preserve the current SPA report as a deliverable, or redesign around copilot-generated briefs?
4. Which outputs are internal-only vs prospect-safe?
5. What is the approval flow before a report can be shared externally?

Copilot UX

1. What does "chat-first" mean operationally: all actions initiated in chat, or chat plus persistent audit canvas?
2. Should users see module cards/tabs, or should the copilot summarize module state unless asked?
3. How should screenshots and browser evidence be shown inside chat?
4. What controls are needed to prevent the copilot from using low-confidence data in outbound copy?
5. Should the copilot proactively recommend next best action after factcheck?

Skill-to-SaaS Conversion

1. Which current skills are highest-priority to fully port next?
2. Should markdown scratchpads remain as first-class artifacts, or become generated views over structured records?
3. How much of audit-data.json should be preserved exactly for backward compatibility?
4. Should deterministic scripts be reused directly, wrapped, or rewritten as internal services?
5. What is the canonical module output contract for each skill after migration?

Data / Evidence

1. What is the database schema for claim registry, citations, source labels, confidence, and factcheck state?
2. Where should screenshots and rendered reports be stored?
3. How long should account intelligence cache live by module?
4. How should degraded-mode outputs be represented in the UI?
5. Should evidence tier rules be enforced at write time, render time, or publish time?

Enterprise / Auth

1. What identity provider should be used for company rollout?
2. What roles exist: viewer, runner, editor, approver, admin?
3. Are audits account-private, team-shared, or company-wide by default?
4. What audit logs/compliance artifacts are required?
5. Do users need tenant separation by region/team?

Future Components

1. Where should Scout sit: shared fetch layer, module-specific tool, or independent service?
2. Which Scout integration has highest ROI after baseline: competitor newsroom, IR/PDF, or browser audit support?
3. When does Discovery OS enter the product: after report generation, as AE call prep, or as the primary output?
4. How should Discovery OS hypotheses be validated against factcheck gates?