Skip to content

Packages for working AI outcomes.

Start with one qualified problem. Discovery decides whether the next step is a workshop, retainer, first outcome, or enterprise proof.

Packages are how clients buy outcomes. Products are the reusable capabilities used to deliver them.
Entry points

Start with a qualified next step.

Qualify one target, test fit in a workshop, then choose the package and operating mode that make the outcome safe to ship.

Qualification gate

Discovery Call

Free / 15 minutes

A short call on one named business problem: a vendor question, a workflow to fix, a report to trust, or a decision surface to clarify.

  • Qualified yes, no, or not-yet answer
  • Workshop recommendation when there is fit
  • Source material and owner needed for the workshop
  • Enterprise proof path when a workshop is too small
Focused workshop

Outcome Workshop

Free or low-cost / 2 hours

A live or Cairn-mediated working session on a qualified target. The goal is to decide what should be bought next, not to turn qualification into open-ended implementation.

  • Outcome hypothesis
  • Source and evidence reality check
  • Recommended next package
  • Retainer fit signal
  • Written assessment of what was found
Starting retainer

AI Advisory and Outcomes Retainer

Starts at $8k/month

A recurring cadence for teams that need AI decisions, outcome priorities, and small follow-through work to keep moving between larger builds. The retainer keeps context, receipts, and next-step decisions current without making activity the product.

  • Regular AI advisory tied to live work
  • Prioritized outcome backlog and next-step decisions
  • Small bounded work when the result fits the cadence
  • Receipts and context carried into the next scope
  • Larger outcomes scoped separately when they need their own sprint
Outcome packages

Buy the outcome, then choose how it runs.

Each package is scoped by what the client needs to accomplish, what evidence must survive review, and how much delivery responsibility CMD+RVL should carry.

Fast proof

First Result Sprint

One spend line, workflow, report, or risk surface the business can evaluate quickly.

Leaves behind
  • Working result
  • Source trail
  • Acceptance test
  • Next-scope map
Spend pressure

Vendor Review / Workflow Sprint

Pressure-testing a known vendor, manual process, or fragile internal workflow before a procurement or operating decision.

Leaves behind
  • Working workflow
  • Coverage comparison
  • Adoption path
  • Decision readiness checklist
Before modernization

Decision Evidence Sprint

Teams that need to understand how decisions are made before automating or rebuilding the surrounding process.

Leaves behind
  • Source inventory
  • Decision map
  • Evidence model
  • Reusable metadata layer
Local data

Office Data Product Sprint

Office or department files that need to become a structured data product, such as thousands of Excel files feeding comps or review workflows.

Leaves behind
  • Schema
  • Extraction pipeline
  • Database or tables
  • Provenance and QA workflow
Recurring state

Operated Outcome / Monitor

Decision-critical state that needs cadence, drift detection, alerts, webhook/API delivery, and provenance records.

Leaves behind
  • Recurring delivery
  • Health signals
  • Delivery ledger
  • Correction path
Client environment

Client-Owned Outcome Setup

Sensitive workflows that need security review, client-side operation, or a scoped path to long-term internal ownership.

Leaves behind
  • Evidence model
  • Deployment assessment
  • Reference workflow
  • Support boundary
Partner bench

Channel Package Enablement

Implementation partners that have buyer access and delivery capacity but need CMD+RVL-supported methodology, evidence patterns, and QA boundaries.

Leaves behind
  • Package template
  • Delivery playbook
  • Evidence standards
  • Co-delivery and QA model
Operating modes

The same outcome can run in different places.

Choose the outcome first, then choose the operating mode. Mode answers where data lives, who hosts and operates the workflow, and whether the proprietary CMD+RVL stack is involved.

CMD+RVL-operated

CMD+RVL hosts, processes, and operates the workflow in our environment, with the outcome and receipt delivered back to the client.

Fit

Fastest path when the client can share approved data and wants the lowest internal IT lift.

Tradeoff

Requires an approved data-transfer or processing boundary before client data moves into CMD+RVL-operated systems.

Client-side CMD+RVL stack

Where feasible and approved, the CMD+RVL stack is scoped for a client-approved cloud, Databricks, Azure, or on-prem boundary.

Fit

Best when sensitive data, internal adoption, or long-term client-side operation matters.

Tradeoff

Requires a Foundation Build, stack license, support terms, security review, and deployment coordination.

Open-source / off-the-shelf delivery

The outcome is delivered with approved standard tools and no proprietary CMD+RVL stack operating the workflow.

Fit

Useful when procurement simplicity, client tool policy, or partner-led implementation matters most.

Tradeoff

Usually slower and less reusable because CMD+RVL memory layers are not operating the workflow; CMD+RVL still defines the receipt, QA gates, and acceptance standard where we are involved.

Site model

How Packages, Products, and Outcomes fit.

This is the mental model the rest of the site should reinforce.

Packages

The commercial and delivery shapes: workshop, retainer, sprint, operated monitor, client-owned outcome setup, or channel enablement.

Outcomes

The result CMD+RVL is trying to produce and defend: the workflow, monitor, report, data product, or decision surface the business actually needs.

Products

Reusable CMD+RVL capabilities and assets used inside packages, including Signals, Data Products, Outcome API / Feature Feeds, and monitors.

Operating Modes

Where and how the work runs: CMD+RVL-operated, client-side CMD+RVL stack, or open-source/off-the-shelf delivery.

Product surfaces used inside packages

Pricing FAQ

Common package questions.

These are the boundaries a buyer, legal team, or delivery partner usually needs before the first scope.

What is the difference between packages and products?

Packages are how clients buy and operate work with CMD+RVL. Products are reusable capabilities and assets, such as Signals, Data Products, Outcome API / Feature Feeds, and monitors, that can power those packages.

Where should a new client start?

Most clients should start with a free 15-minute discovery call on one concrete problem. If there is a fit, the next step is a free or low-cost two-hour workshop that produces an outcome hypothesis, recommended package, and retainer fit signal.

What is included in the AI Advisory and Outcomes Retainer?

The retainer starts at $8k/month and is outcome-scoped, not time-scoped. It gives the team a recurring AI advisory cadence, a prioritized outcome backlog, small bounded follow-through work when scope allows, and receipts and context that carry into the next decision. Larger outcomes are scoped separately when they need their own sprint.

Is enterprise proof work the normal starting path?

No. Enterprise proof work is an exception for prospects whose expected scope is too large for a two-hour workshop to qualify. Most prospects should use discovery call and workshop before deciding on a retainer or first outcome.

Do all packages require CMD+RVL to host the work?

No. The package is chosen separately from the operating mode. The same outcome can be CMD+RVL-operated, scoped for a client-approved environment where feasible, or delivered with open-source/off-the-shelf tools when that is the right procurement path.

How are larger packages priced?

Scoped packages are priced after discovery based on the outcome, data hydration path, delivery cadence, data sensitivity, operating mode, hosting or operation boundary, support level, and any required stack licensing.

Can partners or subcontractors deliver under this model?

Yes. In channel delivery, CMD+RVL can provide the methodology, receipt standard, architecture, QA gates, and acceptance rubric while a partner or subcontractor carries more implementation work under an agreed package model.

If you can name the first result, we can usually name the package.

Bring the vendor, workflow, report, file pile, or decision surface that matters now. The first conversation is about choosing the smallest useful result and the operating mode that makes it safe to ship.