B2A: Why Your Next Customer Might Be an AI Agent, Not a Human

B2A: Why Your Next Customer Might Be an AI Agent, Not a Human

For decades, businesses have organized around two customer relationships: B2B (business-to-business) and B2C (business-to-consumer). A new category is now emerging alongside them B2A, or Business-to-Agent and it's forcing companies to rethink who, or what, they're actually building for.

What Is B2A?

B2A describes a business model where companies design products, websites, and infrastructure primarily for AI agents to interact with, rather than for human eyes alone. Instead of a person browsing a webpage, comparing prices, and clicking "buy," an autonomous AI agent does that work on the person's behalf — researching, negotiating, and completing transactions.

The term is credited to Y Combinator, whose "request for startups" called for companies that provide services directly to AI agents rather than to the humans behind them drawing a parallel to how algorithmic trading already dominates stock exchanges [1]. A widely shared essay later popularized the concept, describing AI agents as "always-on power users" that expect clean APIs, structured data, and instant decisions a framing that stuck across the industry [1], [2].

Why It's Happening Now

Two things are converging to make B2A real rather than theoretical:

1. AI agents are getting genuinely capable. They can now maintain context across multi-step tasks, understand nuanced intent, and act autonomously booking travel, restocking inventory, or purchasing software without a human clicking through every step.

2. Commerce infrastructure is adapting to meet them. OpenAI and Stripe's Instant Checkout lets ChatGPT users complete purchases without leaving the conversation, reportedly reaching hundreds of millions of weekly users [6]. Google has piloted agentic checkout inside Gemini, and Perplexity has added instant purchasing to its search results [6]. Payment networks are moving too Visa, Mastercard, and PayPal have all demonstrated agent-ready APIs [2].

The consumer appetite backs this up: research from Salesforce found that roughly a third of consumers now say they'd rather buy digitally through an AI agent than through direct human interaction, and nearly 40% are comfortable letting an agent schedule appointments for them [5].

From UX to AX (Agent Experience)

Traditional UX design is built around human psychology emotion, visual hierarchy, persuasion. B2A introduces a parallel discipline sometimes called AX (Agent Experience), which prioritizes logic, structured data, and clear permissioning instead.

The practical difference matters: a human might forgive a confusing layout or a broken link and figure it out anyway. An agent generally won't it will simply error out and move to a competitor with better documentation [8]. In an AX-first world, API documentation starts to matter as much as, or more than, visual design [8].

What Businesses Actually Need to Do

Across the commentary on B2A, a few concrete requirements keep coming up:

• Structured, machine-readable data. Product, pricing, and availability information exposed through formats like JSON-LD and well-documented APIs, so agents can parse and act on it without guesswork [5].
• Reliable, documented APIs. Not just for humans building integrations, but as the primary "storefront" an agent interacts with [8].
• Support for emerging agent protocols. A stack of open standards is forming to handle different layers of agent interaction [10]:
• MCP (Model Context Protocol) — Anthropic's protocol for connecting agents to external tools, data, and context.
• A2A (Agent-to-Agent Protocol) — Google's protocol for letting agents from different vendors discover each other's capabilities and coordinate tasks.
• ACP (Agentic Commerce Protocol) — developed by OpenAI and Stripe to handle the handshake between chat-based commerce surfaces and merchant payment systems.
• AP2 — a protocol layer focused on cryptographically signed, autonomous payments with audit trails.
• UCP (Universal Commerce Protocol) — a standard co-developed by Google and Shopify to connect agent discovery to actual checkout, letting merchants expose catalogs and inventory to A2A-compatible agents.
• These protocols aren't competitors so much as different layers of the same stack most production systems end up combining two or three of them rather than picking just one [10].

The Business Case — and the Risk of Waiting

Analysts see this as a genuine distribution shift, not a passing trend. Some industry commentary frames B2A as an imminent channel that could reshape category market share within the next couple of years, arguing that brands optimizing for agents today will control "the invisible shelf" tomorrow much as early B2C ecommerce players captured lasting first-mover advantage [2], [9].

Enterprise consultancies frame it more structurally: this isn't a plug-in feature or an "AI adoption" checkbox, but a deeper shift in how architecture, data, platforms, and governance need to be organized so a business is legible, interoperable, and actionable by intelligent agents without losing strategic coherence [4].

A Practical Starting Point

For teams not ready to overhaul their entire stack, a sensible sequence looks like:

1. Expose structured data first — clean, accurate JSON-LD or API-based product/service data.
2. Add MCP support — the foundational layer that lets agents pull context and use tools reliably.
3. Layer in A2A — once you need agents to negotiate or coordinate with other agents, not just consume your data.
4. Adopt commerce-specific protocols (ACP/AP2/UCP) as needed — once you're ready to accept agent-initiated payments, with the audit and trust guarantees that require.

The Bottom Line

B2A doesn't replace B2B or B2C — it adds a new layer on top of both. The businesses paying attention now are treating AI agents the way early web businesses treated search engine crawlers: as an audience that needs to be able to find, understand, and transact with them programmatically, not just visually. The companies that get their data structured and their APIs agent-ready early are positioned to become the default choice when an agent goes looking the ones that don't may simply become invisible to a growing share of their own customers.

References

[1] M. Cunningham, "The rise of Business-to-Agent (B2A) companies," Medium, Feb. 4, 2025. [Online]. Available: https://medium.com/@mcunningham1440/the-rise-of-business-to-agent-b2a-companies-ad0f1e063f47. [Accessed: Aug. 24, 2026].

[2] "B2A: The business-to-agent model and the next retail disruption," Kantar. [Online]. Available: https://www.kantar.com/north-america/inspiration/retail/b2a-the-business-to-agent-model-and-the-next-retail-disruption. [Accessed: Aug. 24, 2026].

[3] "Understanding Agent-Driven Commerce and AI-Powered Buying," Kibo Commerce, Jun. 16, 2026. [Online]. Available: https://kibocommerce.com/blog/understanding-agent-driven-commerce/. [Accessed: Aug. 24, 2026].

[4] "Beyond B2B and B2C: Preparing for the business-to-agents (B2A) era," Thoughtworks, May 27, 2026. [Online]. Available: https://www.thoughtworks.com/insights/articles/preparing-for-the-business-to-agents-b2a-era. [Accessed: Aug. 24, 2026].

[5] "The rise of Business-to-Agent (B2A): why optimizing for AI Agents matters," Welcome Spaces, Mar. 6, 2025. [Online]. Available: https://www.welcomespaces.io/blog/the-rise-of-business-to-agent-b2a-why-optimizing-for-ai-agents-matters. [Accessed: Aug. 24, 2026].

[6] "Business-to-Agent: When Your Next Customer Isn't Human," IBM iX, Dec. 10, 2025. [Online]. Available: https://ibmix.de/en/blog/business-to-agent-when-your-next-customer-isnt-human/. [Accessed: Aug. 24, 2026].

[7] "What is B2A (Business-to-Agent)? And how to optimize for it," Drimify. [Online]. Available: https://drimify.com/en/resources/b2a-businesstoagent-optimize/. [Accessed: Aug. 24, 2026].

[8] "B2A (Business-to-Agent): Designing for Non-Human Users," Product Leaders Day India, Dec. 5, 2025. [Online]. Available: https://productleadersdayindia.org/blogs/agentic-ai-product-management/designing-for-ai-agents-b2a.html. [Accessed: Aug. 24, 2026].

[9] "B2A: The business-to-agent model and the next retail disruption," Kantar Retail IQ. [Online]. Available: https://kriq.kantarretailiq.com/en/p-iq/insights/blogs/b2a-the-business-to-agent-model-and-the-next-retail-disruption. [Accessed: Aug. 24, 2026].

[10] "MCP vs A2A vs ACP vs AP2: Agentic Commerce Protocols Compared (2026)," Cipher Projects, Jul. 8, 2026. [Online]. Available: https://www.cipherprojects.com/blog/posts/agentic-commerce-protocols-mcp-a2a-acp-ap2-compared/. [Accessed: Aug. 24, 2026].