ARCA
The Agentic Marketing Framework for AI Business Transformation and Enterprise AI Architecture. Developed by Rohit Prabhakar as his thesis, based on testing parts of these ideas across Fortune 50 companies, including Visa, McKesson, Thomson Reuters, and FIS.
Five stages. Five layers. One compounding system designed for agentic transformation at a Fortune 50 scale.
Where is your commercial organization actually?
The free AI marketing maturity model and agentic AI maturity model. Five levels. Six dimensions. The enterprise AI diagnostic that powers the Assess stage of ARCA. Roughly 60 percent of Fortune 500 marketing functions are at Level 2 today. Most believe they are at Level 3.
L1 Fragmented The Tool User No enterprise impact. No AI business transformation underway. | L2 Accumulating The Tool Library ~60% of Fortune 500 sit here. Tools without enterprise AI architecture. | Turning Point L3 Connected Connected Enterprise Marketing Fewer than 15% of Fortune 500. First agentic transformation begins. | L4 Orchestrated Agent-Led Growth Engine Fewer than 5% of Fortune 500. Multi-agent enterprise AI architecture operational. | L5 Compounding The Commercial Moat Year-3 lead competitors structurally cannot close. | |
|---|---|---|---|---|---|
| Context and Memory | Brand voice, customer details, and prior decisions live in heads, decks, and Slack. Marketers re-explain context every time. | Brand kits and prompt libraries exist as files. They sit in storage. They do not flow into the work. | Brand, customer, and decision context loads into every workflow automatically. The data is clean, traceable, and consistent. The system, not the marketer, holds the memory. | Context updates as the world changes, not on a quarterly refresh. | Years of decisions and outcomes are a strategic asset. Searchable, contextual, protected. |
| Customer Intelligence | Customers are addressed as segments. Personalization is first names. The individual is invisible. | Account-based marketing works well. Segment-level personalization is reliable. Individual-level is a roadmap line item. | Each customer is known as one person across channels, with their permission. Agents pull this data through stable connections, with an audit trail. | Every interaction is shaped by who the customer is, what they have done, and the moment they are in. | Real-time individual intelligence at global scale, with privacy intact. |
| Orchestration | A human moves work between every stage by hand. AI helps at single steps. Nothing connects to anything. | AI helps individuals at point steps. Multiple disconnected tools across teams. No shared platform. | At least one full marketing workflow is rebuilt around agents and runs end-to-end. The function operates on a shared agent platform. | Multiple agents work together across content, audience, campaigns, and service. Work moves 10 to 15 times faster. | The same engine extends across marketing, sales, service, and product. The customer experience runs on one system. |
| Governance and Trust | Brand and legal review happens after the fact. No audit trail. No clear list of what could go wrong. | Policy documents exist. Whether they are followed depends on which reviewer catches it. | Brand, compliance, and quality checks are built into the workflow. The risks are named. Every agent has a defined identity and access. | Governance runs as a layer in the system. Drift is detected, not discovered. | Trust is the moat. Audit trails, eval results, and explainability are published commitments. |
| Operating Model | No one owns AI outcomes. Roles unchanged. No oversight. Adoption is voluntary. | An AI lead or center of excellence exists, but marketing owns AI alone. | Marketing, IT, finance, and legal share clear ownership. Reskilling is funded. Each agent has a named owner and review cadence. | The CMO is the orchestrator. Roles are redesigned to manage agents. | Marketing, sales, service, and finance run on the same data, metrics, and P&L. |
| Learning and Compounding | Every campaign starts from a blank page. What worked last time is forgotten. | Wins are written up in retrospectives and live in slide decks. | Outcomes are measured inside the system that produced them. What worked feeds the next campaign automatically. | Every approval and rejection trains the system. Each decision makes the next one better. | Every win and loss is captured, owned, and fed back to make the agents smarter. The loop runs on the institution. |
Why most enterprise AI is not moving the revenue needle.
Roughly eight in ten companies have deployed AI. Roughly eight in ten report no measurable bottom-line impact. The cause is not model capability or budget. It is architecture. AI business transformation fails not because of the model, but because of the missing enterprise AI architecture underneath it.
ARCA is the agentic marketing framework developed as Rohit Prabhakar's thesis based on testing across Fortune 50 companies. It moves an organization from AI as a personal tool to AI as a structural moat.
Four stages. One compounding system.
Five agent layers. Six dimensions. One coherent enterprise AI architecture.
The ARCA Framework maps five distinct agent layers to six commercial dimensions, creating an enterprise AI architecture that gets smarter every quarter.
Signal Agents, Context and Memory: Brand, ICP, decisions, and competitive intelligence load into every workflow.
Insight Agents, Customer Intelligence: Identity-resolved, consented data. Approval and rejection signals train the system continuously.
Action Agents, Orchestration: Multi-agent workflows across content, audience, campaigns, and service.
Guardian Agents, Governance and Trust: Every agent action audited, explainable, reversible.
Orchestration Agents, Operating Model: Cross-functional accountability between marketing, IT, finance, and legal.
The architecture that gets smarter every quarter.
Most enterprise AI deployments depreciate over time. Models go stale. Adoption plateaus. Lessons evaporate when the team turns over. ARCA is built to do the opposite.
Every customer interaction feeds the data layer. Richer data raises agent intelligence. Smarter agents produce better customer experience. Better experience generates more revenue and richer signal. The loop runs on the institution, not on individuals.
This is why the gap between a function at Level 3 of the Maturity Model and a function at Level 5 is roughly 24 months and structurally hard for competitors to close. They are not just ahead. They are accelerating.
ARCA reads differently from each seat at the table.
compounding growth engine.
The hardest part of this transition is not technical. It is convincing the organization that marketing now runs as a system, not as a series of campaigns.
A shared architecture.
The CMO who shows up with architecture, not asks, is the CMO the CIO can build with.
not a cost centre.
A direct line from technology spend to revenue impact, with reporting that arrives at the cadence Finance actually needs.
not a marketing project.
The question is not whether your organization deploys agentic AI. The question is whether you build the architecture that makes it compound.
ARCA is the engine.
Market-of-One is the destination.
Every customer is a market. That is the Market-of-One thesis: the age of the segment is over and the era of the individual has begun.
Believing it is not enough. To treat every customer as a market of one at enterprise scale, across 200 countries, across millions of interactions, requires an architecture that makes it operationally real.
ARCA is that architecture. It is the system that takes the Market-of-One thesis from manifesto to machine. Assess where you are. Architect the agent ecosystem. Command the 90-day deployment. Amplify the compounding flywheel.
Read the Market-of-One ManifestoReady to build the
growth architecture?
Rohit works with a small number of organizations each year on ARCA implementation, executive advisory, and transformation leadership. If you are building the commercial organization for the AI era, start here.
Common questions.
The ARCA Framework is an agentic marketing framework and enterprise AI architecture developed by Rohit Prabhakar as his thesis, based on testing parts of these ideas across Fortune 50 companies including Visa, McKesson, Thomson Reuters, and FIS. ARCA stands for Agentic Revenue and Customer Experience Architecture. It connects AI business transformation to measurable revenue through four stages: Assess, Architect, Command, and Amplify.
ARCA stands for Agentic Revenue and Customer Experience Architecture. It is the AI-driven revenue architecture designed to move an organization from AI as a personal productivity tool to AI as a structural commercial moat, one that compounds and gets smarter every quarter it runs.
Most AI frameworks are theoretical or vendor-built. ARCA is an open agentic marketing framework developed by Rohit Prabhakar as his thesis based on testing parts of these ideas at Fortune 50 companies including Visa, McKesson, Thomson Reuters, and FIS. It is practitioner-developed from over $1 billion in real results, not consulting theory.
The ARCA Command stage is a 90-day production deployment with ROI reporting at 30, 60, and 90 days. Assessment takes 2 to 4 weeks. Architecture design takes 4 to 6 weeks. The full ARCA cycle is designed to deliver measurable AI business transformation results within the first quarter.
The Commercial AI Maturity Model is the AI marketing maturity model and agentic AI maturity model that powers the Assess stage of ARCA. It grades organizations across five levels, from Fragmented at Level 1 to Compounding at Level 5, across six dimensions. Free diagnostic available. No login required.
In-Flow AI is the principle that intelligence should be delivered inside the workflow where the decision happens, with no context switching required. It is the core operating principle of the ARCA Architect stage and the reason ARCA-informed deployments achieve higher adoption rates.
ARCA is designed for CMOs, CDOs, CIOs, CFOs, and CEOs at enterprise organizations deploying AI across commercial functions. It is built for co-ownership across the C-suite. It is most relevant for Fortune 500 organizations where AI investment has not yet produced measurable P&L impact.
Rohit Prabhakar tested parts of the ideas behind the ARCA Framework at four Fortune 50 companies: Visa, McKesson, Thomson Reuters, and FIS, spanning financial services, healthcare, and professional services. Combined, his work at these companies generated over $1 billion in measurable business value and formed the basis of the ARCA thesis.