Brief

Procurement guide: buying context supply chain work

Written for the procurement reader: what you are buying, how it is priced, what you own afterwards.

Published Revised

Procurement for enterprise AI context work should buy an operable supply chain, not a longer pilot. Evaluate vendors on whether they can map sources, structure, semantics, validation, retrieval, agent operations, and the learning loop against your real systems—and whether they can state what will remain out of scope in the first engagement.

Scope a first step that produces decisions, not demos. The interview maps the schema landscape; a half-day Diagnostic Briefing within ten days of it turns that map into written findings and a buy/no-buy. A two-week Agent Context Cost & Failure Audit is a practical entry point: it uses telemetry and traces from one named agent workflow to show where spend and waste concentrate and where the agent fails on wrong context, without promising savings. A three-to-four-week Core Model Blueprint then assesses the connectors, ranks gaps by production impact and sequences the Core Model with a first implementation quarter. Managed operations should be priced only after ownership, write-back rules, and success measures are clear.

Reject proposals that jump from workshop to "agent platform" without a production path: access controls, audit, runbooks, and a named owner for each stage of the chain. Prefer firms that treat maturity as levels you can verify—ad hoc through self-improving, not as marketing stages. Prefer firms that design engagements so pilots can graduate into governed operations rather than restart as science projects.

What is being bought

Assessment engagements buy a written deliverable: a diagnosis, a maturity placement per stage with evidence attached, and where the Blueprint phase is included, a sequenced plan with effort estimates. Implementation engagements buy working software delivered on a quarterly milestone.

Pricing basis

Engagement ladder and pricing basis
EngagementFeeBasis
Diagnostic Briefing$4,500Half day, within ten days of the interview, fixed
Agent Context Cost & Failure Audit$12,0002 weeks, one named agent workflow, fixed
Core Model Blueprint$32,5003 to 4 weeks, one named agent workflow, fixed
Implementation$75,000First quarter, fixed: three systems, one Core Model, the MCP endpoint, CoreModels licence included; expansion quarters $60,000 – $180,000 by scope
Managed Operations$9,000 – $18,000Per month, quarterly rolling: Care $9,000 · Ops + Eval $18,000

Every fee above is fixed before the engagement starts. The Core Model Blueprint is internally split into Diagnose at $15,000 and Blueprint at $17,500, and the second half can be declined once the first is delivered. Implementation’s first quarter is a fixed package; expansion quarters are priced by scope and fixed at quarter start. Managed Operations is two SKUs: Care, and Ops + Eval, which adds the evaluation harness on schedule, incident attribution and on-call for the named agent workflows.

Ownership and exit

  • Reports, blueprints, runbooks and code are owned outright by the client
  • A blueprint we produce can be executed by your team or another vendor
  • No engagement creates a dependency on continued Ariesnet involvement

Questions

Answered directly

How is the commercial relationship structured?

Engagements are scoped and confirmed in an engagement letter. Standard master terms are available; specific clauses are negotiated during scoping. Fees follow the published ladder. Nothing beyond the signed letter is committed until both parties execute that document.

How is scope controlled across rungs?

Each rung has fixed deliverables and duration. Work outside those boundaries is out of scope unless added by written change order. Later rungs start only after prior deliverables are accepted and a new engagement letter is signed.

How are change orders handled?

Material scope, timeline, or fee changes require a written change order before work proceeds. Estimates for added work are provided during scoping discussion and confirmed in the engagement letter or change order, not assumed from prior rungs.

What are typical payment terms?

Payment schedule and invoicing norms are stated in the engagement letter and confirmed at contracting. Standard practice is milestone or time-bound billing aligned to the rung. Exact net days and deposits are set in that letter, not assumed from this page.

How is client data handled?

Access is limited to what the engagement requires. Data handling, retention, and return or destruction are defined in the engagement letter and any required data-processing addendum. Client systems are not modified beyond the agreed scope.

How does an engagement start?

The interview, then a written fixed-scope fixed-fee proposal. We recommend starting at the Agent Context Cost & Failure Audit — it is approvable without a steering committee. Nothing requires you to enter at the bottom of the ladder — a Diagnostic Briefing within ten days of the interview is common when the funding decision is contested.

What system access do you need?

Assessment work is read-only: usage telemetry, a content sample, and time with the people who own the sources. Implementation access is scoped explicitly at the start of each quarter and is limited to the systems in that quarter's scope.

Who owns the deliverables?

You do. Reports, blueprints, runbooks and code produced in an engagement are yours outright. A blueprint you commission from us can be executed by your own engineers or by another partner — that is a deliberate property of how the ladder is priced, not a concession.

Terms used

Definitions

Context supply chain
The end-to-end path that sources, structures, validates, retrieves, and governs enterprise context so agents can act on material that is correct, current, and usable. See also: Maturity level, Retrieval quality
Maturity level (L0 to L5)
A six-point placement applied per stage: L0 Ad hoc, L1 Structured, L2 Semantic, L3 Validated, L4 Operational, L5 Self-improving. Stages are scored independently — a mature retrieval layer sitting on L0 sources is a common and diagnostic pattern. See also: Context supply chain

Related engagement

Want this assessed against your own estate?

The Core Model Blueprint takes one named agent workflow through all seven stages with your systems and your content, and ends in a sequenced Core Model your team could execute without us.