Budget Clarity AI brings natural-language, conversational access to your Budget Clarity data — right inside the AI assistant your team already uses. Financial users can ask Claude, ChatGPT, or Google Gemini everyday questions like "What's the total requested amount for the FY26 Budget Plan?" or "Which budgets are over their approved amount?" and get specific, data-backed answers grounded in live Budget Clarity records — not general knowledge or guesswork.

Built on Proven Optics AI Core's Model Context Protocol (MCP) framework, Budget Clarity AI exposes 18 purpose-built tools covering entity discovery, financial summaries and spread analysis, ownership and approval tracking, and direct record lookups. Every query runs under the requesting user's own ServiceNow session, so results are automatically scoped to that user's existing Budget Clarity permissions — no data elevation, no separate access model to manage.

Because it's read-only by design, Budget Clarity AI lets financial teams get faster answers on spend, approvals, burn rate, and ownership without any risk to underlying budget data. Setup is centralized: a single role grants end-to-end access, and administrators retain full control over which AI clients can connect through AI Core's OAuth-secured, admin-approved connection workflow.

**Key benefits:**

- Get instant, accurate answers on budgets, approvals, and spend — without leaving your AI assistant
- Results are always scoped to what the requesting user is authorized to see
- No manual data pulls or cross-referencing multiple budget records
- Read-only architecture — zero risk of unintended changes to budget data
- Works with Claude, ChatGPT, and Google Gemini Enterprise out of the box

- **Natural-language Q&A over live Budget Clarity data** — financial users ask budget questions in Claude, ChatGPT, or Gemini and get answers grounded in real, current Budget Clarity records rather than general knowledge.
- **18 purpose-built tools across four categories** — discovery (entity resolution, session reference trees), financial (totals, spread data, burn rate, over-budget detection), ownership (individual and group responsibility), and general-purpose (record lookups, direct platform links, user context).
- **User-scoped, ACL-enforced data access** — every tool query runs under the invoking user's own ServiceNow session, so results reflect exactly what that user is authorized to see. No elevated service account and no data leakage across entitlement boundaries.
- **Entity resolution via partial, multi-keyword matching** — users can reference budgets, sessions, and line items by approximate name rather than needing exact matches or sys\_ids.
- **Single-role access model** — one role (x\_prop\_pob\_ai.user) bundles Budget Clarity read access, AI Core framework access, and Enablement utilities, avoiding manual multi-role assignment.
- **Group-aware ownership resolution** — correctly attributes budget responsibility whether it's held individually or through group membership, addressing a common gap in ad hoc reporting.
- **Structured error handling** — tool failures return structured output to the calling AI client instead of hard failures, so the assistant can still respond and note data unavailability.
- **Built on the AI Core MCP framework** — inherits OAuth 2.1 transport, client approval workflow, and audit logging without Budget Clarity AI needing to implement any of it independently.