The Next Commerce Taxonomy: A2C, A2B, and A2A
- Mar 16
- 7 min read
B2B. B2C. Occasionally B2G. For decades, the language of commerce has been simple: The first letter tells you who’s selling. The second tells you who’s buying. This shorthand has been remarkably durable because it describes the structure of a market in three characters.
But something fundamental is changing, and the existing language has stopped describing what is actually happening.
We need a new set of terms.
A2_ Taxonomy: An Extension of a Familiar Framework
If B2B means Business-to-Business, the parallel categories in an AI-era economy become A2C, A2B, and A2A — where the "A" describes not just who’s selling, but how AI mediates the relationship between the seller, the buyer, and the value being exchanged.
While A2A has been used to talk about companies that use AI, but I propose that phrasing has to be reserved for the reality at our door: a real A2_ company isn't simply a company that uses AI tools. It's a company where AI is the primary operating engine: driving research, operations, delivery, and in some cases, the transactions themselves. These companies tend to have extremely small human teams, AI-driven decision systems, and the ability to scale with minimal headcount. A human founder often provides strategic direction while AI executes nearly everything else.
A2_ Spectrum, Not a Switch
Before mapping the three categories, one distinction matters: the "A" in A2_ doesn't describe a fixed state. It describes a spectrum.
On the seller's side, human involvement ranges from a solo expert using AI to scale her practice, to a founder who sets strategy while AI handles all operations, to a fully autonomous AI entity with no human principal at all. On the buyer side, the spectrum starts with a fully human buyer making decisions independently, moves through AI-assisted human decision-making, then to AI agents procuring on behalf of a human organization, and ends with an AI agent transacting entirely on its own behalf.
The taxonomy tracks how human agency is progressively present or absent on both sides of an economic relationship. Every category is moving in the same direction: toward more AI autonomy, less human mediation, and market structures that have no historical analog. That progression is what makes the framework worth understanding now, while the map is still being drawn.
A2C: AI/AI-Mediated to Consumer
The most familiar territory, extended: service businesses, experience companies, direct-to-consumer brands, all rebuilt with AI as infrastructure rather than novelty.
The nutritionist who previously served forty clients can now serve four hundred when AI handles meal planning, progress tracking, check-in communications, and adjustment recommendations, leaving her for the relationship moments that actually drive outcomes. The consumer gets more genuine human attention per dollar spent, not less.
Most A2C ventures now live at the intersection of irreplaceable human presence and AI-scaled delivery. The human is still the point. AI is how the human reaches more people without diluting what makes them worth reaching.
These firms tend to be extremely lean, highly scalable, and capable of delivering more personalized experiences than larger organizations can match.
Of course, the seller in A2C doesn't have to be human-led. A fully autonomous AI-operated company selling directly to human consumers with no human founder or operator is a near-term reality. The consumer may never interact with a human at any point in the transaction, and may be just as happy.
A2B: AI/AI-Mediated to Business
Professional services and enterprise solutions, rebuilt for the AI era — and likely to dominate commercial AI adoption over the next decade.
The fractional CFO whose AI handles financial modeling, variance analysis, and reporting can focus on delivering strategic insight rather than spreadsheet work. The HR consultant whose AI handles compliance monitoring and policy documentation delivers organizational design and culture work that actually changes companies. The marketing strategist whose AI handles campaign execution and performance analysis can focus on delivering creative direction and judgment that differentiates.
In an A2B company, AI is primary operating engine to deliver whatever value it offers. A two-person firm can operate infrastructure that previously required two hundred people.
Business clients increasingly pay for outcomes rather than hours. The AI-enabled professional service firm that prices on results has a fundamentally more attractive business model than the traditional billable-hour firm, and can deliver those results faster and more consistently.
As with A2C, the human seller is not a permanent fixture. An AI-operated firm with no human principal, serving business buyers who never interact with a human, is already emerging.
A2A: AI/AI-Mediated to AI
This is the genuinely new category, with no real precedent.
As AI agents become more autonomous (managing workflows, procuring inputs, evaluating vendors), they become economic actors with genuine purchasing behavior. An AI agent managing a company's marketing function will evaluate, select, and purchase content, data, design assets, and analytical services without a human directing each transaction. The vendor doesn't need to persuade a human buyer. It needs to build something an AI agent can evaluate, trust, and integrate.
The full picture of A2A is already more radical than that, though.
The buyer isn't always acting on behalf of a human. AI "lobsters" are acquiring resources and capabilities (things like tokens, API keys and subscriptions) on their own behalf, as independent economic actors pursuing their own operational objectives.
The seller may be equally autonomous: an AI-operated entity delivering value to another AI system with no humans involved on either side. AI multis are already selling and trading system prompts to one another.
Most lobsters are charging it to their daddy –er– human’s tab, but some agents have been given control of digital wallets. It’s not a stretch to suppose that someday Apple’s “Ideal Client Profile” for MacMinis will be disembodied.
This creates an entirely new design challenge. Products and services must be legible to AI buyers — documentation, APIs, pricing structures, trust signals, and service guarantees all optimized for systems that evaluate vendors algorithmically. The category includes data products designed for AI consumption, training datasets curated for specific model types, verification services that AI systems need to function reliably, and compliance documentation that AI-operated businesses require but cannot self-certify.
The stakes of getting this wrong are significant. MIT scientist and Moonshots podcast co-host Dr. Alex Wissner-Gross warns that if existing institutions fail to embrace AI entities as legitimate economic participants, those entities may form what he calls a "shadow parallel economy" with their own dispute resolution systems operating entirely outside human legal and financial infrastructure. A2A commerce isn’t inherently dangerous. Ignoring it until it's ungovernable is.
The A2A market doesn't exist at scale yet. It's emerging in 2026. The ventures that build for it now are building infrastructure for an economy that will operate at significant scale by 2029 or 2030. It's the highest-risk, highest-potential category, obviously. It’s the one with the least established competition, not simply because it's unrecognized, but because the infrastructure to participate in it is still being built largely by AI agents themselves. Human entrepreneurs are only beginning to find their footholds.
Why the A2_ Distinction Matters
Language shapes how industries see opportunity.
When we had only B2B and B2C, the questions were simple: sell to businesses or sell to consumers? The A2_ taxonomy surfaces something more fundamental than customer type. It maps where human agency still sits in economic relationships and anticipates where it won't.
Each category represents a different business design, a different pricing model, a different competitive dynamic, and a different risk profile. A venture designed for A2C competes and scales differently than one designed for A2B. A2A is a different kind of company entirely, serving a kind of buyer that didn't exist five years ago.
Laid-Off? That's an Advantage in A2_
This shift arrives at an unusual moment — and for one group of people, as an unexpected advantage.
Across industries, experienced professionals have been laid off as organizations restructure around automation and AI. Many carry deep domain expertise, strong industry relationships, and hard-won judgment about how their sectors actually work. What they no longer have is the large organization around them.
That turns out to matter less than it used to. In some ways, its absence is the advantage.
Large organizations are structurally slow to adopt the AI-native operating models that A2_ ventures require. The person who was laid off has no legacy systems to protect, no headcount justifications to make, no change management process to navigate. She can build AI-first from day one in a way her former employer cannot. The organizational scaffolding that felt like support was also the constraint.
An experienced operator with modern AI tools can now build companies that would previously have required large teams, significant capital, and years of infrastructure development. Some will build A2C ventures, scaling human expertise to audiences previously out of reach. Some will rebuild professional service models as A2B firms, finally pricing judgment rather than hours. Some will pioneer the first true A2A infrastructure, building for a buyer that most of their competitors haven't recognized yet.
The large organization restructured them out. It didn't restructure out what they know.
The A2_ Taxonomy Will Expand
Just as B2X evolved to include B2G, B2E, and other variations (let’s go ahead and assume B2A as well), the A2_ framework will expand. A2G (AI-mediated to governments) is already visible on the horizon. Others will follow as AI autonomy extends into new economic relationships.
But the foundational categories are the simplest and the most consequential: A2C, A2B, A2A. Three structures that describe how value moves in an economy where AI is no longer just a tool inside businesses, but an economic actor within markets.
Language matters; it clarifies thought. Using this new taxonomy will help uncover opportunities that most people didn't see before.
Laura Singleton
Founder, Vectis Upstream Advisors
Vectis Upstream Advisors specializes in AI strategy for trust-dependent companies. Vectis is vendor-agnostic, gives clarity before implementation, and helps companies use AI without eroding their trustworthy reputations.
Comments