valueIQ: The Glass-Box Experiment

The first vendor to publish its pricing as a machine-executable model — and a test of whether transparency alone is a pricing strategy.

Editorial disclosure. valueIQ is was co-founded by Steven Forth, one of this book's authors. Michael Mansard is an investor. It is included because it is, as of this writing, the reference implementation of machine-readable pricing and value models. Where the model falls short of the book's own counsel, the case says so.

Case Facts

CNI: Credit Native Index

CGVS: Credit Governance Visibility Score

In One Breath

valueIQ — an AI Value Sales Agent and Pricing Intelligence Agent — prices with a credit-subscription hybrid: four tiers, workspace-pooled credits at a constant three cents, one-month capped rollover, and a published weight table that meters objects held (products, deals, seats) rather than actions taken. It is the first vendor to publish both its value model and its pricing model as machine-readable JSON conforming to an open standard, making it the most agent-evaluable pricing in the market. The glass box also reveals what is missing: an undisclosed top-up price, an unconstrained right to reprice, thin mid-sequence governance.

1. The Business and the Buyer

valueIQ.ai launched in early 2026. Its flagship Value Sales Agent generates CFO-ready business cases, competitor pricing teardowns, and deal coaching from deal context — no CRM integration required — built on the firm's value-modeling methodology from 100+ pricing engagements (valueiq.ai pricing FAQ). A companion Pricing Intelligence Agent produces structured competitor pricing analyses (pricing.valueiq.ai).

The declared segment is B2B SaaS and agentic-AI companies with sales-led motions: $5M–$200M ARR, 5–200 quota-carrying reps, $25K–$150K average deal sizes, where deals need finance sign-off (valueiq.ai FAQ). The published value model names the reference alternatives precisely: a dedicated value engineer (~$200K/year), a $50K+ consulting engagement, 10–15 days of manual spreadsheet work per business case, or a generic LLM whose numbers a finance committee will reject (valueiq.ai value model; valueiq.ai homepage).

The archetype is Domain Oracle: bounded domain (value quantification and pricing intelligence), proprietary methodology, a growing data moat (2,000+ pricing pages ingested), and attribution that is achievable because the product's own output is an attribution artifact. There is a Process Automator overlay — it explicitly substitutes for value-engineering labor — and, through The Value Project open standard, an unmistakable Trust Infrastructure ambition.

2. The Model As Built

Foundation (C1–C4). The unit is an abstract credit — no tokens, no compute metrics. Four tiers: Free ($0, 500 credits/month), Professional ($120/month, 4,000), Teams ($960/month, 32,000), Enterprise (custom, flexible commitments) (valueiq.ai pricing page, accessed 2026-08-05). Billing is monthly only; no annual plan is defined ("commitment_months: 1, adjustment: 0%" in the published model). The published weight table has five consumption rules (valueiq.ai pricing model, effective 2026-05-20):

Two structural facts stand out. First, the price per credit is constant at $0.030 across both paid tiers — the volume curve is flat until a discretionary Enterprise discount. Second, four of the five rules are monthly holding charges on objects, not per-action charges. The co-pilot conversation itself — the generation work — is unmetered; you pay to hold a product model, a live deal, a seat. Seats are priced inside the credit system: a team member is simply 1,000 credits a month. The Free tier's 500 credits are exactly one standard product (300) plus one active deal (200) — packaging arithmetic that is clearly deliberate.

Lifecycle (C5–C13). Overage on paid tiers is pay-as-you-go via a top-up add-on — whose price is not disclosed, in-app or in the JSON ("amount: null... Contact sales"). Rollover is one month, capped, on all tiers including Free; unused credits then expire. Credits pool at the workspace level and are not assigned to individuals. No refunds for unused credits — published as a hard rule. Draw order, transfer, and cancellation treatment of outstanding credits are not documented. The agent warns on low balance.

Governance and GTM (C14–C17). Here is the case's reason for existing: valueIQ publishes its full pricing model as a PricingModel JSON instance conforming to The Value Project standard (PMS v1.4), and its value model — ten value drivers with executable equations and benchmark ranges that candidly label seven of nine improvement claims as vendor_estimate — alongside it, with an MCP server and REST API for programmatic access. The pricing page links "machine-readable pricing model (for AI agents)" directly. Against that, one sentence sits unrepaired in the terms: "valueIQ reserves the right to adjust credit costs and package sizes" — with no notice period, no grandfathering, no published change protocol.

3. The Design Audit

Two staircase chains deserve development. First, the C1 decision to meter objects held rather than actions taken makes burn predictable — which is why C5 overage rarely bites and the lifecycle feels benign. But the same choice decouples credit burn from value moments: a deal that closes for $150K burned the same $6/month as a deal that dies. Nothing downstream can retrofit that linkage; the outcome connection must eventually be re-established at C1/C2, exactly as the DSM predicts. Second, the C8 dodge propagates upward against the company's own thesis: the most machine-readable pricing model in the market contains a null where the top-up price should be, which means the buying agents valueIQ is evangelizing for cannot compute a customer's worst-case monthly cost. The glass box makes the gap more visible, not less.

The C14 finding is the sharpest irony. Chapter 14's build sequence runs balance visibility → admin analytics → forecasting → alerts → audit granularity → then publish machine-readably. valueIQ executed step six first, brilliantly, while steps two through five remain undocumented. Each step earns the trust the next one spends; publication is currently spending trust the middle of the sequence has not yet earned.

4. The Value Test

The published value model makes the EVE reconstruction unusually easy — the vendor did it for us, in schema. For the core segment (B2B SaaS, 5–200 reps): reference value is a $200K/year value engineer, a $50K consulting engagement, or 10–15 rep-days per manual business case. Positive differentiation spans nine recurring drivers — competitive win-rate lift (median claim +5%), no-decision conversion, cycle compression, 90% business-case labor reduction, VE headcount avoidance, orphaned-deal preservation, discount reduction (median claim 15%) — mapped cleanly to the valueIQ/Nagle value-driver categories. Negative differentiation is modest: this is a human-in-the-loop tool, so the autonomy risk discount is small — and in a reflexive twist, the model monetizes the discount, claiming value for reducing the risk that a generic AI's unauditable ROI numbers get a deal rejected by finance (driver 8). The autonomy risk appears on the value side of the ledger.

Now the two dials. Dial one, price per credit, is $0.030 everywhere. Dial two, the weights, produce effective prices of $6/month per active deal and $0.90 per CFO-ready presentation — against deals worth $25K–$150K ACV and a claimed labor saving of thousands of dollars per business case. Teams at $11,520/year is 5.8% of the single-VE reference value, before counting any of the nine recurring drivers (analyst calculation from published figures). Even applying the vendor's own minimum benchmark claims, the implied VtB is not "above 10" — it is plausibly above 100 (analyst estimate). This is penetration pricing for the land phase, and it is honest to call it that. But by the book's own logic the weights track objects and simplicity, not value: a CWAS-style check finds the value-dense action (the deal) carrying one of the lightest weights. The dials are calibrated for adoption, not capture. The model is value-documented to an exemplary degree, and value-priced to a deliberately modest one — a gap that will have to close as the segment-level evidence accumulates.

5. What the Market Did Next

The market record is early and thin — itself a data point; no independent G2 or community reviews were found as of August 2026, consistent with a product roughly one year old. What exists: vendor-reported traction of 2,000+ pricing pages analyzed and 1,000+ value cases generated; a named customer testimonial crediting the tool with justifying pricing and increasing sale size (Bill Mastin, CEO, Cadient); and pricing stability — the tier structure and rates observed in the May 10, 2026 archive snapshot are unchanged in August 2026.

The louder reaction is to the transparency play, not the price. The Value Project launch (May 2026) drew ecosystem engagement — open schemas, Apache 2.0, a published roadmap toward foundation governance, and collaboration with Michael Mansard. An independent AI-led audit of the schemas judged them "a very solid Level 1–2 transparency framework... but not yet a full Level 3–4 standard" — parseable and executable, with provenance and evaluation contracts still implicit. And the credits skeptics have engaged directly: an April 2026 Impact Pricing debate challenged whether credits "break the connection between price and value" at all. valueIQ's object-holding weights are, in effect, its answer: make the credit legible enough and the abstraction objection weakens.

Scorecard

CNI: Credit Native Index

CGVS: Credit Governance Visibility Score

6. Lessons for the Practitioner

1. Metering objects instead of actions is a legitimate answer to taxi-meter anxiety — but it defers value capture. Holding charges make burn predictable and conversations free; they also make a won deal and a dead deal cost the same. The outcome linkage must eventually be rebuilt upstream, at C1/C2, where the DSM says it lives.

2. A constant price per credit buys legibility with expansion economics. A flat $0.030 with no volume curve is easy to explain and easy for an agent to compute — and surrenders the commitment and volume levers that make credit businesses forecastable.

3. Machine-readable pricing audits the publisher first. A null where the top-up price belongs is invisible on a marketing page and glaring in JSON. Publish nothing you are not prepared to have an agent interrogate.

4. "We reserve the right to adjust credit costs at any time" is the sentence the governance chapter exists to delete. A published change protocol — notice period, grandfathering, weight-change log — would cost little and complete the trust architecture the publication step has already promised.

5. Pricing at six percent of the reference value is a land strategy, and land strategies expire. With a published value model claiming a VtB far above ten, the question is not whether to raise capture but which dial to turn, for which segment, and with what governance — before the market answers it for you.
   Pull quote: The glass box does not just let buyers see in. It shows the vendor, with uncomfortable precision, exactly which parts of the pricing system have not been built yet.