THE VALUE PROJECT · OPEN STANDARD

Five open JSON schemas make value models and pricing models machine-readable, so AI buying agents and selling agents can retrieve, instantiate, exchange, and verify ROI programmatically.

Steven Forth, Co-Founder, valueIQ · August 13, 2026

QUICK ANSWER

Five JSON schemas (ValueModel, CustomerVariables, PricingModel, DealConfiguration, InvoiceStatement) are now available under open licence at github.com/The-Value-Project. They let any vendor publish machine-readable value models and pricing models, so AI buying agents and selling agents can retrieve, instantiate, exchange, and verify ROI programmatically.

A pricing page tells you what a product costs. A value page tells you what it's worth.

Most B2B vendors have a pricing page. The format is predictable: tiers, features, seat counts, add-ons. Readable by humans. Parseable by machines. An AI agent can scrape a pricing page and extract the pricing model without human intervention.

Value models have lived in spreadsheets and slide decks. Unreadable by any system that did not create them. No standard format. No machine-parseable structure. No way for an AI buying agent to verify a vendor's ROI claim without a human reading a PDF and deciding whether to believe it.

The value page standard changes that.

Two repositories. Five schemas. All implementation-ready.

In the value-models repository:

In the pricing-models repository:

The chain: the model defines, the variables instantiate, the configuration selects, the statement computes.

How the five schemas compute Value Surplus

VALUE

ValueModel → CustomerVariables → Total Economic Value

PRICE

PricingModel → DealConfiguration → InvoiceStatement → Total Price

VALUE SURPLUS

Total Economic Value − Total Price

The machine-verifiable gap that makes a price defensible.

A ValueModel says: "Our product delivers value through three drivers: time savings, error reduction, compliance risk mitigation. Here are the equations. Here are the benchmark ranges."

CustomerVariables says: "This customer processes 50,000 transactions per month. Their current error rate is 4.2%. Their compliance penalty exposure is $2M annually."

The model runs. The output: Total Economic Value for this specific customer.

The PricingModel says: "Professional tier, $X per seat per month, add-ons priced at Y% of base."

DealConfiguration says: "This customer selected Professional, 200 seats, two add-ons."

InvoiceStatement says: "Total Price = $Z annually."

Total Economic Value minus Total Price. The machine-verifiable gap that makes a price defensible.

If a vendor claims their product delivers $500K in annual value and charges $100K, the Value Surplus is $400K. An AI buying agent can retrieve the ValueModel, instantiate it with the buyer's own variables, verify the $500K claim, and confirm the $400K surplus exists.

If the math does not hold, the claim is rejected. No human reads the vendor's slide deck. The agent evaluates the model and moves on.

This is not an afterthought. Agent-to-agent exchange is what the schemas are built for.

Schemas are Apache 2.0. Specifications are CC BY 4.0.

Apache 2.0 applies to the five JSON schemas, the machine-readable format definitions. Anyone can implement them, extend them, embed them in commercial products. No permission required. No royalty.

CC BY 4.0 applies to the specification documents, the human-readable explanations of what the schemas mean, how they should be used, and what the design principles are. Attribution required. No other restriction.

Do not blur the two. The code is Apache. The documentation is Creative Commons.

Governance is Stage 1. valueIQ leads. Every decision is made in public GitHub issues with a fourteen-day comment period. The roadmap is visible. The changelog is public. At ten outside implementations, governance moves to an independent technical steering committee with one seat per contributor organization and no single-entity veto. At broad adoption, the standard moves to the Linux Foundation or equivalent neutral governance body.

The standard is built to outlive its sponsor.

A buying agent and a selling agent can exchange value models programmatically.

SCENARIO

A CFO asks their AI buying agent to evaluate three vendors. The agent retrieves each vendor's ValueModel, instantiates it with the CFO's company variables (deal count, average ACV, current win rate, current discount rate), and computes Total Economic Value for each option. It retrieves each vendor's PricingModel, applies the company's seat count and selected tier, and computes Total Price. It calculates Value Surplus for all three. It ranks them.

No human read a business case PDF. No human copied numbers from a slide deck into a spreadsheet. The agent verified every claim programmatically.

The schemas that make this exchange possible exist today. The public infrastructure to make every vendor's model retrievable in one place is what we are building next. The schema is the foundation it stands on.

Clone the repositories. Read the specifications. Implement the schemas in your product.

If you sell B2B software, publish a ValueModel alongside your pricing page. Define your value drivers. Cite your equations. Source your benchmarks. Let an AI agent verify your ROI claim instead of forcing a human to trust it.

If you build procurement software, implement schema support so your users' buying agents can retrieve and evaluate value models from every vendor they consider.

If you build sales tools, generate ValueModel instances from deal context so your users can publish machine-readable business cases.

The schemas are open. The governance is transparent. The use cases are immediate.

Do I need valueIQ to use these schemas?

No. The schemas are open and vendor-neutral. Any product can implement them. valueIQ implements them because we generate schema-compliant output on every business case and pricing analysis. But the standard does not require our platform.

What licence applies to what?

Apache 2.0 applies to the five JSON schemas in the repositories. CC BY 4.0 applies to the specification documents. If you implement the schemas in your product, Apache 2.0 governs. If you write documentation explaining how to use them, CC BY 4.0 governs.

Is this a standard or a product?

Both. It is an open standard that any vendor can implement. valueIQ is the first implementation. We generate schema-compliant output on every business case and pricing analysis. But the standard is designed to support any implementation, not just ours.

How do I contribute?

File an issue in the relevant repository. Propose a schema extension. Comment on open proposals. At ten outside implementations, governance opens to a technical steering committee with seats for all contributor organizations.

Build on the standard

The schemas are open. Clone them, implement them, and publish a value model an agent can read.