ARX Messaging Framework

Audience: founder-led sales, CEO conversations, solutions engineering, proposal drafting, website copy, and anyone shaping ARX narrative.

Purpose: capture the messaging, positioning, and proposal logic we developed so ARX speaks with one voice across pitches, memos, and the proposal generator.

Core Thesis

ARX is AI workforce infrastructure.

The job is not to help a company "use AI more." The job is to help a company turn AI into aligned, auditable, proactive execution inside the work that matters most.

The best short form:

> Horizontal AI is smart at scale. ARX aligns execution at scale.

The more complete version:

> Horizontal AI makes intelligence broadly available. ARX turns that intelligence into proactive execution aligned to the organization’s objectives.

Category Language

Lead with:

• AI workforce infrastructure
• the operating model for AI labor
• proactive AI execution aligned to company objectives

Do not lead with:

• "control plane"
• "AI governance platform"
• "agent security"
• "marketplace"

Those can all be supporting ideas, but they are weaker category openers.

The Core Contrast

ARX should not attack horizontal AI as useless. That is not credible and not necessary.

The correct contrast is:

> Horizontal AI is smart at scale. ARX is aligned at scale.

Supporting versions:

• Horizontal AI helps employees. ARX deploys execution.
• Horizontal AI improves workflows. ARX staffs them.
• Horizontal AI responds to prompts. ARX executes against objectives.
• Horizontal AI is smart automation. ARX is governed digital labor.

What this means:

• Horizontal AI can search, reason, and automate broadly.
• ARX is for work that must stay aligned to changing priorities, reporting lines, controls, and management intent.

The key internal line:

> Being smart is not the same as being aligned.

Objective-Driven Positioning

ARX should always frame the work as objective-driven, not prompt-driven.

Use:

> ARX agents execute against objectives, not prompts.

Or:

> ARX aligns AI execution to the organization’s objectives, then keeps it operating proactively inside the company’s real workflows.

Important implication:

• the value is not "better AI responses"
• the value is measurable leverage for knowledge workers
• the end state is employees doing more with less

The best articulation:

> The real enterprise question is not how to build agents. It is how to deploy proactive AI execution that measurably increases knowledge-worker output as priorities change.

Workforce Performance Thesis

Do not frame ARX as fixing broken teams.

Frame it as raising the operating level of the workforce.

Use:

> ARX raises the operating level of the workforce.

Or:

> ARX lifts the floor, the median, and the ceiling of workforce performance.

Meaning:

• lower-leverage work becomes more consistent
• average performers gain throughput
• top performers gain the leverage of a larger team

This is the meaning behind the earlier shorthand of:

• C -> B
• B -> A
• A -> superhuman

External version:

> ARX helps every employee operate at a higher level.

First Principles At The Employee Layer

One of the strongest internal insights is:

> Most AI systems are designed from the platform inward. ARX starts at the employee layer and reasons outward.

This gives us:

> ARX raises the operating level of the workforce by applying first principles at the employee level.

Meaning:

• start from the actual role
• start from the actual objective
• start from the recurring work, bottlenecks, handoffs, and decisions
• then decide what AI should own, what stays human, and how the work is supervised

Outcome-First Rule

ARX should always lead with the business result, not the category.

The structure is:

1. business outcome
2. named wedge
3. timeframe
4. mechanism

Use:

> We can deliver [specific business outcome] in [named workflow / market / function] within [timeframe].

Then:

> ARX makes that result operational, governable, and defensible by aligning AI execution to scoped roles, named supervision, and customer-verifiable records.

Never lead with:

• "we built a control plane"
• "we are an AI governance platform"
• "we help companies scale AI"

Those are category statements, not decision triggers.

Company-Language Rule

ARX should speak in the customer’s language, not ours.

That means:

• use the company’s own business units
• use the company’s own titles
• use the company’s public commitments
• use the company’s current financial or operating pressure

Do not write:

• growth
• execution
• intelligence
• control

unless those are clearly internal framing devices only.

Instead write:

• Market Access
• Trust & Safety
• Investor Relations
• Customer Operations
• Regional GM, Puerto Rico
• Director, Clinical Operations

The rule:

> ARX operationalizes the exact functions and roles that carry the company’s objectives into execution.

Proposal Standard

The right benchmark is the Lilly and Coinbase memo quality standard:

• detailed enough for a CEO to make a decision
• specific enough for a CFO to challenge the math
• concrete enough for an operator to imagine implementation

The proposal is not a brochure. It is an operating proposal.

Every good ARX proposal should have:

1. one named business objective
2. one named operating wedge
3. one defensible topline number
4. three or four mechanism components behind that number
5. company-language roles or operating cohorts
6. explicit regulatory posture
7. a real execution plan
8. a bounded ask tied to the customer’s calendar

Proposal Archetypes

We now have at least two memo shapes.

1. Function / wedge memo

Use when the pitch is to:

• a GM
• a BU leader
• a revenue owner
• a market owner
• a clinical, commercial, or operating leader

Characteristics:

• one book or wedge
• twelve-month value window
• 3-part number decomposition
• 3-5 specific roles
• 90-day scoped pilot

Examples:

• Lilly Puerto Rico
• EviCore
• SME operator memos

2. Company transformation memo

Use when the pitch is to:

• president / COO / P&COO
• CEO-adjacent operator
• transformation owner
• company-wide operating sponsor

Characteristics:

• company-wide operating leverage
• public commitment or letter as the anchor
• sequenced implementation thesis
• AI-native pods across quarters
• broader ROI ladder
• explicit "what breaks the math"

Example:

• Coinbase

Proposal Generator Standard

The generator should stay simple on input:

• company name
• LinkedIn company profile

But it should produce decision-grade output.

That means it should infer and assemble:

• company type
• revenue-base band
• likely buyer type
• latest news
• 10-K / 10-Q / investor language where relevant
• public commitment or current operating pressure
• objective hypothesis
• workflow wedge
• regulatory regimes
• company-language roles and titles
• topline number
• mechanism decomposition
• execution logic

The generator should feel like:

> simple intake, deep memo

not:

> simple intake, generic sales copy

Proposal Generator Sections

For wedge / function proposals

The generator should be able to produce:

• Cover block
• Opening
• About ARX
• The moment
• The number
• Where the numbers come from
• Roles we would operationalize
• Compliance and operating boundaries
• Plan of execution
• Ask

For company transformation proposals

The generator should be able to produce:

• Cover block
• Opening
• About ARX
• The implementation thesis
• The number
• Where the numbers come from
• The operating model
• Regulatory / audit perimeter
• Plan of execution
• Return on investment
• Ask

Role Logic

When the proposal is role-based, every role must be tied directly to the inferred objective.

Each role block should show:

• role or function name in company language
• business unit
• financial boost range
• objective tie
• role-specific mechanism
• research AI agent
• production AI agent
• coordination AI agent
• operating outcome line

The rule:

> A role only belongs in the proposal if it clearly advances the named company objective.

Mechanism Math

Topline claims should never float without mechanism.

Each component should read like:

• mechanism
• why it is believable
• how it converts into business impact
• where the estimate is conservative

Examples:

• capacity expansion
• cycle-time compression
• access or launch acceleration
• cost-per-output reduction
• audit-readiness compression
• incident-velocity compression
• renewal or retention protection

Key trust-builder:

> Explicitly say what is not included in the headline when that increases credibility.

Compliance Positioning

Compliance should always be framed as tighter, not looser.

Use:

> Compliance posture tightens, not loosens.

And:

> ARX does not see the underlying model traffic. ARX sits beside the work and gates it.

Proposal rule:

• name the actual regimes
• name the actual regulators when relevant
• say the data stays inside the customer environment
• say each AI agent has fixed scope, credentialed identity, and a verifiable log

The Ask

The ask should feel like a real operating step, not a sales CTA.

For function / wedge memos:

• typically 30 minutes
• operating owner + budget owner + workflow owner

For transformation memos:

• typically 60 minutes
• recipient + CFO + CTO + one function owner

The ask must state:

• who should be in the room
• what gets decided in that meeting
• what happens if the math does not hold
• what happens if it does
• why the timing matters now

Messaging Guardrails

Do:

• lead with the outcome
• use company language
• tie every role to the objective
• speak to workforce leverage
• show math, not just adjectives
• use public commitments when available
• let the proposal read like an implementation thesis

Do not:

• lead with product category language
• write generic role blocks
• use AI jargon instead of company operating language
• imply replacement in external-facing copy
• overstate precision on the numbers
• write a memo that cannot survive a CFO follow-up

Internal One-Liners

Use these internally and externally when helpful:

• ARX is AI workforce infrastructure.
• Horizontal AI is smart at scale. ARX aligns execution at scale.
• Being smart is not the same as being aligned.
• ARX agents execute against objectives, not prompts.
• ARX raises the operating level of the workforce.
• The real enterprise question is not how to build agents. It is how to increase knowledge-worker output in measurable ways.
• ARX operationalizes the functions that carry the company’s objectives into execution.

Current Standard For /proposal

The root proposal route should now be judged against this standard:

• simple input
• company-specific output
• public-signal grounding
• objective-driven wedge selection
• company-language roles
• defensible number decomposition
• compliance specificity
• real implementation path

If the output does not feel like something a CEO could review and make a decision on, it is not done.