What is PromptKing?
PromptKing governs AI agent behavior, spend, and compliance across Claude, Microsoft 365 Copilot, GitHub Copilot, Google Gemini, AWS Bedrock, and IBM Watsonx — from one system, entirely out of the data path. Every governed run gets a receipt. Every receipt gets a SHA-256 hash. Every CFO gets an answer. Beneath the governance layer runs the Economics Track: the seat-level, subscription-native intelligence this page walks through.
Two tracks, one system. The Governance Track answers what your agents did, whether it should have been allowed, and what the receipt proves. The Economics Track applies FinOps discipline to AI seat licences, token consumption, and compliance risk — giving finance, IT, and the board one place to see, govern, and recover AI spend.
POSITIONING · JULY 2026
Cloud cost platforms own the infrastructure budget — and do it well. But no infrastructure tool has a concept of a ghost seat, a subscription plan tier, or an agent trajectory. PromptKing answers the questions the rest of your stack isn't built to answer: which of your AI subscription seats are wasting budget right now, and what your agents did to cause it.
$210,000
recoverable per year for a 500-seat org at 35% ghost rate
Industry average ghost rate: 35%. The 90-day target: 0% ghost rate.
90 sec
time to flag a ghost seat after connector sync
0 clicks
to generate the CFO Red Report
THE CONFIDENCE LAYER
Every AI FinOps platform shows a green “Connected” badge when an API key validates. That badge tells you the pipe works. It tells you nothing about what kind of truth flows through it. Anthropic exposes authoritative billing. Google Gemini exposes behavioral intelligence — but not per-user billing. IBM Watsonx exposes estimated economics from token counts. PromptKing grades every connector A through D and propagates that grade into every derived metric — so a CFO always knows how much to trust the number they're looking at.
Grade A–D
every connector rated by data quality, not just connectivity
5 conditions
for board-ready certification on any metric
48h
freshness threshold — stale data labeled automatically
Six questions your current AI FinOps tool probably cannot answer:
Vendors expose cost. PromptKing exposes confidence.
ZERO PROMPT VISIBILITY
The question your legal team will ask about any AI governance platform: “Does it need to see our prompts to work?”
Prompt-level attribution is granular but privacy-invasive. It requires reading prompt content to tag costs — which means storing, processing, and potentially exposing the intellectual property inside your prompts.
Session-level attribution maps cost and behavior to business workflows: a session is a use case, a task, a decision. It is the unit your compliance team needs for Article 26 evidence. It is the unit your CFO needs for ROI reporting. It is the unit your IT Director needs for rightsizing decisions.
PromptKing measures AI performance without ever reading AI content. That is not a trade-off. It is a new category.
STACK POSITIONING
PromptKing is not a replacement for your data-governance, cloud-cost, or observability platforms. It is a layer those tools cannot reach.
Microsoft Purview governs Microsoft AI. PromptKing governs everything else.
Cloud cost platforms manage infrastructure spend. PromptKing manages AI subscription intelligence. Both export FOCUS 1.4 conformant data — they are designed to coexist.
Developer observability tools work at the prompt level. PromptKing provides executive and compliance intelligence at the session and outcome level.
The question that positions PromptKing correctly in every enterprise conversation: “What governs the AI your data-governance, cloud-cost, and observability tools cannot see?”
That is where PromptKing operates.
Read: Purview governs your Microsoft AI. PromptKing governs everything else →
THE NORMALIZATION LAYER
Anthropic bills per million tokens. GitHub Copilot bills in AI Credits with a 27× model multiplier. IBM Watsonx bills in Resource Units. Microsoft bills flat per seat. None of them speak the same language. PromptKing normalizes all four into a single per-seat-per-day USD metric — the only number that tells you whether your AI spend is working or wasting.
4 billing architectures
unified to 1 metric
27×
GitHub Copilot Opus 4.7 multiplier — tracked
June 15
programmatic billing split — monitored live
Where PromptKing sits
Every lane below does its own job well. None of them can answer the question in gold. That question is where PromptKing operates — different line item, different buyer, and often the same enterprise needs all of it.
Covers: Compute, storage, network — deploy-to-cost mapping for engineering teams.
Can't see: No seat identity. No subscription plan tiers. No concept of a licence that nobody is using.
“Which of your AI subscription seats is wasting budget right now?”
Covers: Per-call token spend — engineer-queryable, infrastructure-native.
Can't see: No coverage of who holds the licences, who's active, and who left the company three months ago.
“Whose seat is a ghost, and what is the confidence score on that call?”
Covers: One consolidated invoice view across vendors — procurement-grade rollups.
Can't see: Aggregation is not action. Anomalies land in an inbox; remediation happens somewhere else, weeks later.
“When an agent loops, who sees the dollar figure before the bill compounds?”
The FinOps Foundation Mandate
The FinOps Foundation's updated framework mandates ‘Executive Strategy Alignment’ — connecting AI asset tracking to business outcomes and C-suite risk profiles. Our Red Report, Health Scoring, and EU AI Act module fulfill this natively.
Read the FinOps Foundation framework →What the first 60 days are designed to deliver
Modeled on our demo environment:
8 vendors, 1 dashboard — connected in an afternoon.
Found $420/month in recoverable budget in the first sync.
CFO sees the Red Report Monday morning. Calls by noon.
Design target: ghost rate to under 5% in 60 days.
No enterprise contract. No 6-week onboarding. Connect your Anthropic Console or Microsoft Admin Centre and your first insight is live before your next standup.
SOC2 Type II in progress (Q4 2026) · PIPEDA compliant · Canadian data residency · Every engagement starts with a conversation
Already using Apptio, ServiceNow, or Finout?
PromptKing integrates with your existing FinOps stack — it doesn't replace it.
FOCUS 1.4 export live · No custom ETL required