By SOLOMOAT Editorial Team

By Yang Bin, Tencent Research Institute July 16, 2026 | Jiangsu

Yang Bin, Ph.D., is Vice Chairperson of the University Council at Tsinghua University, Professor at the School of Economics and Management, Director of the Center for Leadership Development and Research, and Dean of the Institute for Sustainable Social Value.

💡 Core Strategic Takeaway

• AI agents are becoming active demand-side participants rather than passive tools for human buyers.
• Commercial strategy must optimize for machine-readable value, trust, interoperability, and autonomous procurement.
• B2A changes discovery and competition because agents can continuously compare, negotiate, and execute transactions.

When clients are no longer exclusively human, when competition extends beyond traffic and attention, and as commercial workflows increasingly operate with the "human out of the loop," the established playbooks of the internet era demand unlearning. Professor Yang Bin introduces the concept of "B2A" (Business-to-Agent), providing a structural framework for the commercial arenas emerging in the exponential age of AI. Our mastery of B2B and B2C is complete; B2A has just begun.

Since the internet's popularization, commercial models have bifurcated into Business-to-Business (B2B) and Business-to-Consumer (B2C), categorizing whether products and services are procured by corporate departments or end-users. This binary framework dictates corporate identity and market positioning across venture capital and strategic planning. However, accelerating AI development is rapidly surfacing a new paradigm of critical importance: B2A. The "A" in B2A signifies AI, Agents, or enterprises fundamentally structured around AI as their native operating system, primary productivity driver, and core organizational architecture—entities currently classified as AI-native or "AI-pilled".

Historically, commercial analysis defaulted to a human-centric baseline: transactions occur between humans, companies serve other companies or consumers, and consumers purchase goods and services. Placing AI in the exponent fundamentally alters the base commercial model. AI now generates and consumes an escalating volume of information, data, and user interfaces, while autonomously executing or assisting with a growing share of decisions. The commercial reality is shifting: market participants are no longer exclusively human, and Agents are solidifying into indispensable market actors. Latest data across multiple telemetry sources confirms that by late 2025, or Q1 2026 at the latest, API calls between AI agents surpassed total network requests initiated by human users. The execution perimeter of AI continues to expand beyond conventional projections.

If a procurement Agent can autonomously execute supplier screening, price arbitration, contract negotiation, and fulfillment tracking; if an investment Agent can independently aggregate Analysis, run financial models, and rebalance portfolios; if a household Agent handles the majority of consumer purchasing decisions on behalf of users, the fundamental question arises: is the enterprise still selling to a consumer, or to the Agent acting as their proxy? Is the target of corporate competition human attention, or algorithm permissions, recommendation weights, and scheduling priority? Many executives remain unaware that this new operational arena is scaling rapidly.

While corporate strategy traditionally emphasizes supply-side reform, innovation, and efficiency, the future demands recognizing a structural demand-side revolution. The catalyst is a quiet shift in the entity responsible for expressing demand, screening options, and executing decisions: AI Agents are becoming the proxy and the ultimate decision-maker for escalating market demand. In specific operational environments, the "human out of the loop" architecture represents the defining commercial logic of B2A.

Bottlenecks in traditional commercial workflows consistently stem from the severe limits of human attention, time constraints, and cognitive boundaries. Massive commercial feedback loops frequently stall while waiting for human processing capability. In a B2A ecosystem, commercial operations no longer require human intervention at every node. When human experience, preferences, and decision logic run continuously as an agentic cognitive framework, the operational loop achieves maximum velocity. For the first time, commercial systems can bypass the structural limits of human-in-the-loop design.

A new demand side is forming, introducing novel market entities. B2B and B2C focus on corporate and consumer needs, essentially decoding and serving human requirements. The aggressive expansion of B2A signals that enterprises must engineer services for a non-human entity that acts as the primary driver of demand, decision-making, and value creation. This shift will not eradicate B2B and B2C, much as e-commerce did not extinguish physical retail, but it will rewrite the baseline mechanics of significant economic sectors. B2A is positioned to mature into an independent commercial paradigm, with market scale and growth potential exceeding legacy B2B and B2C sectors. This necessitates a new operational lexicon to serve "A" as a distinct market participant.

Structurally, B2A extends beyond human-to-Agent interfaces to encompass Agent-to-Agent interactions, forming a new commercial topology. As long-horizon planning capabilities and autonomous agency improve, Agent-to-Agent bandwidth, trust protocols, and execution synchronization will unlock new operational dimensions. Contrary to assumptions that B2A belongs strictly to management science and engineering, it will likely spawn a new discipline of Agent Management, running parallel to token economics. B2A is not a uniform utopia; because Agents originate from diverse human operators, they will inherit and generate variations, edge cases, and distinct personalities requiring active management. Enterprises must prepare for the structural friction introduced by "A" as a new market participant. This reality drives the formal introduction of the "B2A" framework.

Skeptics may argue that B2A is merely a segment within B2B or B2C, asserting that businesses and consumers remain the ultimate source of demand and evaluation. This reductive argument severely underprices the qualitative shift triggered by the exponential AI transition. Even viewing B2A as a subordinate layer ultimately guided by humans, it radically alters transaction structures, efficiency, and commercial outcomes. For most enterprises today, capturing B2A requires bridging a cognitive gap, not just a technical one.

The primary hurdle in executing a B2A transition is discarding legacy experience—often requiring a complete operational inversion. The first step into B2A is unlearning the default mechanics of the B2B and B2C eras. Consider this operational scenario: why do service robots in hotels not extend an arm to press elevator buttons? They move between floors seamlessly without physically pressing up/down or floor numbers. This should trigger an "aha moment". Realize that many unquestioned standards of the B2B/B2C eras become counterproductive friction in a B2A model.

Legacy models assume the end-user is human, prioritizing frictionless UI, aesthetic design, and emotional resonance. B2A discards interface aesthetics, optimizing entirely for open APIs, standardized data schemas, and latency stability. Legacy models price human time and attention as scarce commodities, centering the internet and mobile eras on advertising, subscriptions, memberships, and traffic acquisition. For B2A, the actual scarce assets are stable compute and verifiable data. Legacy models assume corporate or consumer clients demand only polished final outputs due to limited attention bandwidth, intentionally discarding decision trees and granular telemetry. B2A systems aggressively ingest this raw operational telemetry; they do not reject unfiltered data. This data retention facilitates precise Agent replication and preserves the potential for algorithmic discovery.

Legacy organizational structures chase growth through internal division of labor and workflow optimization, obsessing over how AI can scale execution. AI-native enterprises prove future organizations require a new operational stack: target alignment, capability distribution, and value sharing. Organizational innovation becomes emergent and generative, relying on ecosystems and decision mechanisms entirely distinct from past models.

B2A requires establishing "carbon-silicon symbiosis" as the baseline assumption. Legacy business models and value creation mechanics face mandatory reconstruction; interaction paradigms across all commercial nodes require disruptive overhaul rather than mere expansion. These shifts demand strategic iteration and rebuilding models from first principles. Dogmas previously viewed as permanent and self-evident must be systematically rewritten today.

The true organizational challenge lies in breaking these dogmas and committing to unlearning. Only by shedding the assumed truths of B2B and B2C, accepting that legacy success pathways are dead ends, and acknowledging the arrival of entirely new business models, service targets, and value metrics, can an enterprise anchor itself in the B2A arena. This allows organizations to rewrite competitive advantages and embrace the infinite game unlocked by exponential AI disruption. Our mastery of B2B and B2C is complete; B2A has just begun.

From inception, operators must actively reject the familiar and unlearn the internet era's successes. This unlearning is particularly critical for the tech giants and top-tier founders who won the previous cycle. These elevated titles recall E.H. Gombrich's distinction in The Story of Art between capital-A "Art" and the ordinary, accessible "artists." The core insight across institutional dynamics remains: for legacy tech titans and established leaders, the ultimate stress test is downgrading from a capitalized institution back to a nimble, lowercase operator.

❓ Frequently Asked Questions

What does B2A mean?

Business-to-Agent describes commercial systems in which AI agents discover, evaluate, negotiate, or purchase products and services on behalf of people or organizations.

How should companies prepare for B2A commerce?

They should make offerings machine-readable, verifiable, interoperable, and accessible through trusted APIs and agent-compatible transaction workflows.

🎓 Deepen Your Strategic Mastery

Turn these strategic insights into a focused, defensible one-person-company operating system with the Garbo Decodes China mini-MBA.

Explore Garbo Decodes China