How we manage, model AI Platforms, AI Models and per-model pricing in ServiceNow

Sanjeev Kumar1
Kilo Sage

Hi All,

We are building an AI catalogue in ServiceNow so employees can request access to AI platforms through Employee Center, and so we can track which models and costs sit behind each platform. Advice on the best-practice data model.

Business scenario

AI Platform (what users request access to): OpenAI, Claude (Anthropic), Azure AI Foundry, Runway, Adobe Firefly, etc.
AI Model (what sits under a platform): for example Claude 3.7 Sonnet, Claude Sonnet 4, GPT-4.1.
Each platform exposes multiple models, and each model has different cost details (for example input/output token price, per-seat or subscription price, per-credit price).
The same model can be reached through more than one route. For example, Claude models are available directly from Anthropic or through Azure AI Foundry, and the price can differ by route.
Users request access at platform level. Finance and the AI governance team need visibility at model and cost level.
What we are trying to achieve

Maintain a master list of providers, platforms, models and rate cards.
Let users request platform access via a catalog item, with manager approval and a follow-up task.
Track cost per model, with a history of price changes.
Report on usage and cost by platform, model and department.

Questions

Should the platform and the model be separate tables with a parent/child relationship, or the same table with a hierarchy?
How should we handle the same model offered through multiple platforms or routes with different prices?
Is fm_contract_rate_card suitable for token-based or consumption-based pricing, or should that be custom?
How do we keep the catalog request (platform level) linked to the model and cost data for reporting?
Is there any out-of-the-box or store solution for tracking AI/LLM vendors and consumption?

0 REPLIES 0