a16z partners Olivia Moore and Justine Moore propose a “B2A economy”
a16z partners Olivia Moore and Justine Moore published an article on November 10, 2026, proposing a “B2A (Business-to-Agent) economy.” The article says that the internet’s main users are no longer humans, but AI agents; as agents shift from answering questions to completing tasks, they will increasingly decide on users’ behalf which products to use, which services to subscribe to, and where to spend.
The article cites data saying that Shopify’s first-quarter AI-driven store traffic increased 8 times year over year, and orders generated by AI search increased by nearly 13 times; HUMAN Security monitored that from August to October, agent browser traffic increased 8 times, with Muse contributing the largest share (60%), and OpenAI products accounting for 30% in total. The article argues that four types of companies stand to benefit: agent models, new products brought about by agents, traditional companies optimized for agents, and agent infrastructure.
Original sources:
Title: Welcome to the B2A Economy
Publisher: Olivia Moore and Justine Moore, Andreessen Horowitz (a16z)
Publication date: 2026-11-10
URL: a16z.com/welcome-to-the-b2a-economy
Source type: official a16z article
Verification status: full page text actually read in this request
Supporting content: proposes the B2A economy; says AI agents have become the internet’s main users and will make decisions on users’ behalf; Shopify Q1 AI traffic rose 8 times and AI search orders rose nearly 13 times; HUMAN Security saw agent browser traffic rise 8 times from August to October; lists four types of companies that may benefit
ABAB AI Insight
The article describes the current stage as the early phase of a platform shift: adoption is already measurable, but the business models it creates have mostly not yet emerged. It places agents alongside past changes in web and mobile internet distribution, noting that AI has changed user capability, but the reshaping of distribution channels has only just begun. The Shopify and HUMAN Security data cited in the article are all publicly available third-party metrics, used to illustrate the scale of growth in AI-related traffic and orders.
The division into four types of benefiting companies is the authors’ judgment: the agent models themselves, products newly created because of agents, existing companies that actively optimize for agents, and infrastructure that supports agent operation. The article uses programming agents as an existing example, citing GitHub and other data to show that agents’ share in code generation is rising, and mentions that companies such as Supabase, Neon, Stripe, and Vercel have reported growth in agent-related usage. These examples are all developer-tool phenomena that have already occurred, rather than results of large-scale consumer adoption.
The article also points out that most companies are currently blocking agent traffic because of fraud and chargeback risks, but the authors believe companies will ultimately need to embrace this channel. It also mentions that ChatGPT has launched Businesses features, allowing companies to claim their business and provide product information. The statements above are all the authors’ analysis of trends, rather than completed transaction-scale or market-share data. Indicators to watch for subsequent progress include the continued pace of growth in AI-related traffic and orders, the number of cases in which agents complete actual payments or subscriptions, and whether companies adjust their policies for blocking agent traffic.
ABAB News · Cognitive Law
As agents shift from answering questions to completing tasks, distribution channels can shift from human-facing to agent-facing.
Growth data for AI-related traffic and orders can be placed alongside rising agent usage shares in the programming field as early signals.
Companies’ current blocking of agent traffic and their future need to win agent business can coexist in the same argument.