Protum Releases VBP-AI 1.0, a Free Open-Source Standard for Pricing Enterprise AI by Outcomes, Not Compute

SINGAPORE - June 16, 2026 - Protum today released version 1.0 of the Value-Based Pricing for AI (VBP-AI) Framework, a free, open-source procurement standard that helps enterprises pay for the economic value an AI system delivers rather than the computational effort it consumes. The framework is vendor-agnostic and available to any organisation under the MIT open-source licence at https://github.com/ProtumAI/vbp-ai-framework.

Software pricing is shifting under the weight of AI. Per-seat licensing is weakening because AI reduces the human hours a task requires, and many AI vendors have moved to token-consumption billing, which charges the buyer for raw compute whether or not the output is correct. Under that model, an organisation pays the same for a verified answer and a confidently fabricated one.

VBP-AI applies a test that enterprises already use with human suppliers. If a consultant gave a board fabricated numbers, the invoice would not be paid. The framework brings that accountability to AI contracts and shifts financial risk back toward the party that controls model quality.

The framework gives sourcing, legal, and technology teams four practical components:

• Three outcome-based billing models, ranging from a fee per successfully completed task, to a fee per verified business resolution, to a share of the cost savings or revenue an AI system generates.
• Three ready-to-use contract clauses to paste into a Master Services Agreement or RFP: no charge for fabricated outputs (intrinsic hallucinations), no charge when an AI agent correctly declines to answer for lack of grounded data, and no charge for the internal reasoning tokens a model generates before producing a final answer.
• An automated "judge gateway" that scores every AI response for factual consistency against its source before the response can be billed. Outputs that meet the agreed accuracy baseline are passed and billable; those that fall below it are blocked and zero-rated.
• A phased sourcing playbook that guides procurement teams through auditing vendors, inserting the clauses into RFPs, deploying the gateway, and reconciling invoices each month.

"If a consultant gave your board fabricated numbers, you would not pay the invoice, and you would not hire them again," said Sandeep Joshi, Founder of Protum. "Yet enterprises pay for AI that does exactly that, because they are billed for compute instead of correctness. VBP-AI lets any organisation write that accountability into its AI contracts. We have released it free and vendor-agnostic because a pricing standard only works if the whole market can adopt it, not only our customers."

Protum has published the full framework on GitHub under the MIT License, which lets any enterprise, procurement group, or industry consortium copy, modify, and distribute it without restriction. Alongside the clause templates and the judge gateway design, the framework includes an implementation playbook for sourcing teams and an evidence base compiled from public AI evaluation leaderboards. The release comes as enterprises face pressure to show returns on record levels of AI infrastructure investment, and as model accuracy continues to vary between systems and shift with each release.

The VBP-AI Framework 1.0 is available now at https://github.com/ProtumAI/vbp-ai-framework and https://protum.ai/vbp.

8 Marina View, Asia Square Tower 1,
Level 42-01, Suite 42041,
Singapore 018960

Protum is the company behind Trusted AI OS, an AI assurance platform built in Singapore for regulated organisations in Asia. Protum helps enterprises map their AI estate to the frameworks their regulators inspect, including the Monetary Authority of Singapore (MAS) FEAT principles, the Infocomm Media Development Authority (IMDA) AI Verify, Singapore's Personal Data Protection Act (PDPA), and ISO 42001, gain runtime visibility into what their AI agents do, and carry a defensible certification mark when they ship. Protum is headquartered at Asia Square Tower 1, Singapore. Learn more at https://protum.ai.

This release was published on openPR.