ARCA Framework Explained: The Agentic Revenue and CX Architecture for Enterprise Leaders

The ARCA Framework exists because a pattern kept repeating itself inside Fortune 50 boardrooms. AI budgets were growing. Pilots were succeeding. And yet, when it came time to report to the board, the numbers were not moving. Revenue was flat. Costs were not declining in any meaningful way. The organization had more AI tools than ever before, and fewer results to show for it.

This is not an isolated observation. McKinsey’s 2025 State of AI report found that while 88% of organizations now use AI in at least one function, only 6% qualify as true AI high performers, companies where more than 5% of EBIT is actually attributable to AI. Gartner is more direct about the consequences: more than 40% of agentic AI projects will be canceled by the end of 2027, primarily due to escalating costs, unclear business value, and inadequate governance.

The ARCA Framework was developed to address this exact problem. Not as a theoretical model. Not as a vendor playbook. But as Rohit Prabhakar’s thesis, built from testing agentic transformation at Visa, McKesson, Thomson Reuters, and FIS, four Fortune 50 companies across financial services, healthcare, and professional services, generating over $1 billion in measurable business value in the process.

This guide explains everything you need to know about the ARCA Framework: what it is, how its four stages work, why the five agent layers matter, and how to use the free Commercial OS Maturity Model to find out where your organization actually stands today.

What Is the ARCA Framework?

The ARCA Framework is an open agentic marketing framework and enterprise AI architecture designed to connect AI investment directly to measurable commercial revenue. ARCA is an acronym for Agentic Revenue and Customer Experience Architecture.

At its core, ARCA answers the question that most enterprise AI programs cannot: why is the AI we are deploying not moving our P&L? The answer, in almost every case, is architecture. Not model capability. Not budget. Not team talent. The architecture that connects AI agents to business outcomes is either missing or broken.

ARCA is published as an open framework, free to use and adapt. It is not a product. It is not a vendor platform. It is a blueprint, built from real deployments at real scale, and shared because the gap between AI investment and AI outcomes is too consequential to leave unsolved.

Why Enterprise AI Is Not Generating Revenue, The Architecture Problem

Before understanding how ARCA solves the problem, you need to understand why the problem exists in the first place. And the data is unambiguous.

These numbers reveal a structural problem that has nothing to do with the quality of AI models available today. The problem is that most enterprise AI deployments are built as tool collections rather than compounding systems. Tools accumulate. Productivity rises in pockets. And then the CFO asks what the AI budget is actually producing, and nobody has a clean answer.

There are three root causes that the ARCA Framework specifically addresses:

1. No shared memory or context. Each AI tool in a typical enterprise stack operates in isolation. It does not know what the others know. Every agent starts from scratch, every time. The result is fragmented intelligence that cannot compound.

2. No governance architecture. Gartner found that only 21% of organizations have a mature governance model for autonomous AI agents. Without it, risk accumulates invisibly until something goes wrong. Most enterprises are building agents before they have defined how to supervise them.

3. No revenue attribution. When AI cannot show a direct line to a measurable business outcome, it becomes a cost center. ARCA is built so that every agent layer maps directly to a revenue or cost metric from day one.

“The companies that will compound through this AI cycle are not the ones moving fastest. They are the ones moving most deliberately, with production-grade governance, scoped pilots, and human-in-the-loop architectures from day one.”

Enterprise AI Agents Adoption Statistics 2026

The Four Stages of the ARCA Framework

ARCA operates across four sequential stages. Each stage builds on the previous one. Each has a specific goal, a specific agent layer, and a specific set of measurable outcomes. Here is how they work.

The Five Agent Layers of the ARCA Framework

ARCA maps the commercial organization to five distinct agent layers. Each layer operates in a specific domain, feeds into the others, and maps to one or more of the six dimensions of the Commercial OS Maturity Model. Together, they form a complete enterprise AI architecture where every part serves the whole.

1. Signal Agents, Context and Memory. Signal Agents are the intelligence layer of the commercial organization. They hold brand voice, ICP data, customer decision history, and competitive intelligence, and they load that context into every workflow automatically. When a marketer starts a new campaign, the Signal Agent already knows the audience, the history, and the guardrails. Context switching drops to near zero.

2. Insight Agents, Customer Intelligence. Insight Agents convert identity-resolved, consented first-party data into individual-level intelligence. They do not work with segments or averages. They work with individuals. Every approval and rejection signal they receive trains the system continuously, meaning the Insight Agent layer gets more accurate every week it operates.

3. Action Agents, Orchestration. Action Agents execute across content, audience, campaigns, and service. They coordinate multi-agent workflows that handle production at a scale no human team could manage manually. This is where the enterprise agentic AI market is most active: multi-agent systems commanded 53.30% market share in 2025 and are growing at 43.50% annually (Mordor Intelligence), precisely because coordinated execution is where AI produces measurable revenue impact.

4. Guardian Agents, Governance and Trust. Guardian Agents continuously evaluate what every other agent is doing. They enforce defined risk taxonomies, run real-time compliance checks, audit every action, and make every agent output explainable and reversible. This is not optional. Gartner forecasts that the Guardian Agents category will capture 10 to 15% of the agentic AI market by 2030, precisely because governance is becoming the primary differentiator between deployments that scale and deployments that get shut down.

5. Orchestration Agents, Operating Model. Orchestration Agents span the C-suite. They are the layer that makes marketing, IT, finance, and legal share a common operating model. They define who owns AI outcomes, how roles evolve, how supervision works, and how the commercial organization adapts as the system matures. This is the layer most enterprises skip, and it is the most common reason AI programs stall at Level 2 of the maturity model.

The Commercial OS Maturity Model: Where Is Your Organization Actually?

The Commercial OS Maturity Model is the AI marketing maturity model that powers the Assess stage of ARCA. It grades organizations across five levels and six dimensions, and it tells a very different story than most self-assessments do.

McKinsey’s 2025 data shows that while 34% of business leaders report deeply transforming with AI, only 6% are genuine high performers. The Commercial OS Maturity Model is calibrated for the version of your organization that exists today, not the version on your roadmap.

The six dimensions graded are: Context and Memory, Customer Intelligence, Orchestration, Governance and Trust, Operating Model, and Learning and Compounding. Each dimension has its own score, its own bottleneck insight, and its own 90-day move.

The full diagnostic is free, available at rohitprabhakar.com/frameworks/arca/maturity-model/, requires no login, and produces a board-ready PDF you can download and share immediately.

ARCA in Practice: What Fortune 50 Results Look Like

The ARCA Framework was not designed at a whiteboard. It was developed through testing at real companies, at real scale, with real accountability for results. Here is what that looks like in practice.

McKesson: $900M in New Revenue Through Individual-Level Personalization

At McKesson, one of the largest healthcare companies in the world, the starting point was Account-Based Marketing. ABM was working. Revenue was moving. The team was confident. And then the question became: what happens when you stop marketing to accounts and start serving the individuals within those accounts?

The CFO evaluating cost savings is not the same person as the supply chain manager worried about distribution reliability. Same account. Three completely different moments of truth. Three completely different conversations required. When the architecture was built to serve individuals rather than accounts, the result was $900 million in new revenue and $40 million in cost savings. Not from a clever campaign. From an architecture that stopped averaging customers and started serving people.

Visa: Personalization at Scale Across 200 Countries

At Visa, the challenge was different. Not one company with thousands of customers. One platform serving 3.9 billion cardholders, hundreds of millions of daily transactions, and over 200 countries. Personalization at that scale is not a marketing problem. It is an architecture problem. The agentic infrastructure tested at Visa demonstrated that individual-level intelligence at global scale is operationally real, not aspirational, when the architecture is built correctly from the start.

Thomson Reuters: 700% Sales Acceleration

At Thomson Reuters, the focus was on the sales and marketing handoff, one of the most consistently broken workflows in B2B commercial organizations. By redesigning the workflow around agents that held shared context, unified data, and real-time customer signals, the organization achieved 700% sales acceleration. Not incremental improvement. A fundamentally different rate of commercial output.

ARCA vs Traditional AI Frameworks: What Makes It Different

The agentic AI market is full of frameworks right now. LangGraph, CrewAI, AutoGen, Semantic Kernel, and dozens of vendor-specific orchestration platforms are competing for enterprise adoption. Understanding how ARCA differs from these approaches is essential for making the right architectural choice.

The most important distinction is not technical. It is philosophical. Most AI frameworks are built to make agents faster or more capable. ARCA is built to make the commercial organization more profitable. That difference in design intent produces fundamentally different outcomes.

Why Governance Is the Real Differentiator in 2026

If there is one lesson from the agentic AI market in 2026, it is this: governance is not a constraint on AI deployment. It is the competitive advantage that determines which deployments compound and which ones get shut down.

Gartner’s 2026 data is explicit: more than 40% of agentic AI projects will be canceled by the end of 2027. The primary drivers are escalating costs, unclear business value, and inadequate risk controls. The organizations that survive this wave are not the ones moving fastest. They are the ones that built governance before they needed it.

ARCA’s Guardian Agent layer addresses this directly. Every agent in an ARCA deployment has a defined identity, defined permissions, and a Guardian design from day one. Every action is audited, explainable, and reversible. Risk is named, contained, and turned into a competitive advantage rather than a liability.

This matters especially for CMOs and CDOs navigating board scrutiny. The commercial leader who can walk into a board meeting with a clear governance architecture, a maturity model showing where the organization stands, and ROI evidence at 30/60/90 days is the leader who keeps the AI budget and gets more of it.

The ROI numbers are compelling. US enterprises project average returns of 192% from agentic AI deployments. But the companies achieving those numbers share one characteristic: they built governance into the architecture before they scaled. The ones failing spent the same money and built the same agents, but skipped the Guardian layer. That is not a technology problem. It is a design problem.

How to Get Started with the ARCA Framework

Starting with ARCA does not require a multi-million dollar platform investment, a dedicated AI team, or a board mandate. It requires honesty about where you actually are, and a clear 90-day move based on that starting point.

Step 1: Take the free Commercial OS Maturity Model diagnostic. It takes 12 questions and five minutes. It will tell you which of the six dimensions is your gating bottleneck, what level you are actually at, and what your specific 90-day move is. No email, no login, no paywall. Available at rohitprabhakar.com/frameworks/arca/maturity-model/.

Step 2: Read the ARCA Framework. The full framework is published openly at rohitprabhakar.com/frameworks/arca/. Understand the five agent layers and which ones your organization has already deployed, even partially. Most organizations are further along than they think in one or two layers, and critically behind in others.

Step 3: Identify your Level 3 unlock. Level 3 of the maturity model is the turning point. It is where AI stops being a personal productivity tool and starts being a system that produces measurable commercial impact. The diagnostic will show you the specific dimension preventing you from reaching Level 3. That is where to focus first.

Step 4: Build governance before you scale. Before adding more agents, tools, or platforms, define your Guardian Agent layer. What can each agent access? Who is accountable for each agent’s outputs? What is the escalation path when something goes wrong? These questions are much easier to answer before you have 50 agents in production than after.

Frequently Asked Questions About the ARCA Framework

The Bottom Line: Architecture Is the Competitive Advantage

The agentic AI market will reach $93.20 billion by 2032. The organizations that win that market will not be the ones with the best AI models or the biggest AI budgets. They will be the ones with the best architecture. The ones that built governance in before they needed it. The ones that mapped every agent to a measurable business outcome from day one. The ones that chose compounding over accumulating.

The ARCA Framework exists because the gap between AI investment and AI outcome is not a technology problem. It never was. It is an architecture problem. And architecture problems have architecture solutions.

The window to get this right is narrowing. Gartner’s data shows that Level 3, the turning point where AI starts producing measurable commercial impact, will become increasingly difficult to reach as Level 4 and Level 5 organizations compound their advantage. The organizations that build the right architecture in 2025 and 2026 will be the ones that are structurally ahead by 2028, in ways their competitors cannot easily close.