Public documentation for governed AI labor
SDKs/Governance/Connectors
Arx / Docs / ARX Value-Based Pricing

Documentation

ARX Value-Based Pricing

Project-Agent-trust-merge / docs/pricing/value-based-pricing.md

Project-Agent-trust-merge product-docs docs/pricing/value-based-pricing.md

Audience: CEO, CHRO, CFO, CISO of an enterprise considering an ARX engagement. Procurement teams reviewing the commercial structure.

One sentence: ARX is a BCG-shaped engagement, paid forward — a startup engagement fee at signing, a performance retainer paid in advance against forecasted deliverables, and a quarterly true-up that flexes the retainer up or down against measured productivity.

---

Why this shape

The buyer for AI Workforce Governance is the CEO, CHRO, and CFO together. That buyer expects an engagement model, not SaaS metering. Productivity gains are denominated in FTE-equivalent hours saved, cost-to-serve deltas, cycle-time reductions — outcomes the buyer already benchmarks against industry peers and against their own prior years.

A SaaS price-per-seat model creates the wrong incentives at this buyer level. Per-seat pricing penalizes the customer for adding capacity exactly when productivity is compounding. The BCG model aligns the commercial relationship with the productivity curve: ARX gets paid more when the customer's productivity gain is larger, less when it underperforms, and the customer always pays forward against a forecast they can budget against.

This is the same shape McKinsey, Bain, BCG, and the larger strategy/operations consultancies have used for decades for transformation engagements. It works because the buyer already knows how to procure against it.

---

Three components

1. Engagement Fee — one-time, upfront at signing

Covers the deployment and the first-cohort onboarding of the digital workforce.

What's inside:

  • Scoping: stage assessment, gap analysis, deliverable schedule.
  • Atlas deployment into the customer's own environment.
  • First cohort of stock agents instantiated against the customer's

existing roster (typically 50–200 agents in the first quarter).

  • Policy bundle install: approval gates, runtime budgets,

framework declarations.

  • Audit chain bootstrap: customer-side S3 bucket + KMS key + first

witness signature.

  • CHRO / CFO / CISO briefings on the operating model, with

documented sign-off.

Reference range, by enterprise headcount:

| Tier | Employees | Engagement Fee | |---|---|---| | 1 | < 2,500 (mid-market) | $500K | | 2 | 2,500 – 10,000 | $1.0M | | 3 | 10,000 – 25,000 | $2.0M | | 4 | 25,000 – 75,000 | $3.5M | | 5 | 75,000+ (Fortune 50) | $5.0M |

The engagement fee is non-refundable once Atlas is deployed and the first cohort is live. Before that point, customer can withdraw with a documented reason and the fee is prorated to the work-in-progress.

2. Performance Retainer — recurring, paid in advance

Quarterly or annual prepayment against the forecasted deliverable schedule. The forecast is generated by the ARX Workforce Modeler, reviewed and signed by the CHRO + CFO before each engagement period.

What's inside the retainer:

  • All stock agents (the 252-cell matrix of Research / Production /

Coordination shapes across 14 functions × 6 levels).

  • Atlas — the customer-private CEO-aide. Bundled, never separately

metered.

  • The intercept API and connector framework.
  • The workforce console (Roster, Personnel files, Manager queue,

Compliance, Modeler).

  • The audit chain and evidence packages.
  • Manager approval routing, including SLA escalation and

Slack/Teams notification.

  • Up to 80 hours per quarter of CISO/CHRO/CFO advisory.

Sizing: the retainer is denominated against the deliverable forecast, not against headcount. A typical mid-market customer's first-year retainer lands at $4M–$8M. A Fortune 50 customer lands at $20M–$30M. Steady-state Year 3 retainers tend to settle 10–15% below Year 1 as deliverables compound and the engagement matures.

The retainer is paid forward at the start of each quarter (or annually, with a 5% discount for annual prepay). Customers paying forward are buying outcomes, not metered consumption.

3. Performance True-up — quarterly, up or down

At the end of each quarter, ARX measures productivity against the forecast in the deliverable schedule. Outperformance generates a success fee; underperformance generates a retainer credit.

What we measure:

  • FTE-equivalent hours saved per cohort (direct comparison against

the forecast).

  • Cost-to-serve delta on the customer's named workflows (e.g.

cost per ticket resolved, cost per deal followup, cost per invoice approved).

  • Cycle-time reduction on the customer's named processes (e.g.

time-to-quote, time-to-merge, time-to-close).

The exact set of measures is locked into the engagement letter at signing, drawn from the modeler's deliverable schedule.

Caps:

  • Outperformance success fee: capped at +25% of the retainer.

Customer cannot be surprised with a runaway bill at quarter end.

  • Underperformance credit: capped at -15% of the retainer.

ARX takes a meaningful but bounded haircut — the customer already paid forward, and a credit returns to them against the next retainer.

True-ups settle within 30 days of quarter end. The measurement methodology is auditable from the customer's own audit chain (witnessed by the customer, never by ARX).

---

What's bundled vs separately billed

Bundled in the retainer:

  • Atlas (CEO-aide, customer-private).
  • All stock agents.
  • Workforce console + intercept API + connector framework.
  • Audit chain + evidence packages.
  • Manager approval routing.
  • 80 hr/quarter of CHRO/CFO/CISO advisory.

Separately billed:

  • Third-party agents from the ARX ecosystem (sold and supported by

their vendors, not ARX). Vendor pricing is the vendor's; ARX collects no margin on third-party agents in the catalog.

  • Connector deep-engineering for non-standard internal systems

(a homegrown ticketing system, an unusual ERP, etc.). Quoted per scope.

  • On-site CHRO/CFO/CISO advisory beyond the included 80 hr/quarter.

Quoted at consulting rates.

---

Worked example — Cisco-shape customer

Customer profile: ~90,000 employees (Tier 5), 6,000 AI-augmented in the first year (the modeler's forecast for "aggressive" risk tolerance, 18-month horizon).

Year 1 commercials:

| Component | Amount | |---|---| | Engagement Fee (Tier 5) | $5.0M | | Performance Retainer (annual) | $24.0M | | Year 1 paid forward at signing | $29.0M | | Projected positive true-up | +$3.0M | | Year 1 expected total | ~$32.0M |

Year 3 steady state:

| Component | Amount | |---|---| | Performance Retainer (annual) | $26.0M | | Expected true-up swing | ±$5.0M | | Year 3 expected total | $26.0M ± $5M |

For the customer, this lands as roughly $80–$100M of cumulative spend over 3 years. Against the forecasted productivity gain (~$300M of FTE-equivalent value at steady state), the engagement returns 3–4× annually. The CFO can model this against named deliverables, not against agent count.

---

Why pay upfront

Same answer as the strategy consultancies: the engagement model is the commercial alignment.

  • Customers paying forward are buying outcomes, not metered

consumption. They commit to the program; the program commits to the deliverables.

  • ARX's incentive on the upside (the +25% success fee) and

downside (the -15% retainer credit) is symmetric and bounded. Both sides know the worst case at signing.

  • The cash-flow shape lets ARX fund the engagement-team capacity

the deployment needs (the 80 hours of advisory per quarter is not free — it's loaded into the retainer).

  • Procurement teams that procure consulting engagements every

year already have a workflow for this shape. SaaS metering forces them to treat ARX as a category they don't have a procurement playbook for; a BCG-shape engagement uses the playbook they already run.

---

Why this is different from a SaaS price card

A SaaS price card is a unit-economics-priced product. The vendor optimizes for ARR per seat, the customer optimizes for seat count. The two parties end up adversarial on capacity expansion: every new seat is a new ARR commitment for the customer.

A BCG-shape engagement is paid against a forecasted deliverable schedule. Capacity expansion is *the engagement compounding*, not a new commitment. Both parties are aligned on the productivity curve: more agents → more deliverables → larger forecast → larger retainer → larger productivity gain. Neither party benefits from the other being squeezed.

This is why ARX never publishes a SaaS price card. The price is denominated in the customer's own outcomes.

---

Forecasting: how the retainer sizing is set

The Workforce Modeler produces, per engagement period, a deliverable schedule:

  • Per cohort: agent count, attached human roles, FTE-equivalent

forecast, cost-to-serve forecast.

  • Per function: rollup of forecasted productivity gain, in dollars

and in FTE-equivalent hours.

  • Per quarter: deployment phasing, when each cohort goes live.

The retainer is sized as a fraction of the forecasted productivity gain — typical engagements price at 25-35% of the dollar-denominated gain. This is the BCG benchmark: a transformation engagement captures a quarter to a third of the value it generates.

The forecast is reviewed and signed by the customer's CHRO and CFO at the start of each engagement period. The true-up at quarter end measures actual against forecast and adjusts.

The forecast is ARX's internal sizing tool, not the customer's contractual commitment. The contractual commitment is the deliverable schedule, denominated in measurable outcomes (FTE-equivalent hours, cost-to-serve delta, cycle-time reduction). Forecast missing the actual is an ARX problem to absorb in the true-up; the customer's commitment is to pay the retainer forward.

---

Operational notes

  • Currency. All commercial figures are USD. International

customers are invoiced in USD via FX-rate-locked engagements at the start of each period.

  • Payment terms. Net-15 from invoice. Engagement fee paid at

signing; retainers paid at quarter or year start; true-up invoices net-30.

  • Renewal. Engagements auto-renew on the same retainer +

forecast at the close of each year, subject to either party giving 90 days' notice of non-renewal. Engagement fee is one-time and not re-charged on renewal.

  • Geographic separation. Each region (US, EU, APAC) has its

own engagement letter and its own audit chain. Cross-region productivity rollups are reported but never commingled.

  • Data residency. Atlas is customer-private. The audit chain

is in the customer's S3 bucket, in the customer's region, encrypted with the customer's KMS keys. ARX never has read access.

---

What this replaces

The legacy volume-tiered per-employee price card (still surfaced under legacy_volume_tiered in the Workforce Modeler's /pricing endpoint for historical compatibility) is being phased out. Engagements signed after Q2 2026 are all on the value-based shape above. The legacy shape is preserved in the modeler so prior engagement letters remain auditable.