Archive Note
Marketing to Robots: How to Win in the Age of the AI Agent
This essay developed the idea that the assistant itself was becoming part of the funnel.
The Channel
The note argued that go-to-market teams were entering a third channel alongside outbound and inbound: agentic discovery. In this channel, the assistant recommends, filters, compares, and frames options before the human buyer may ever visit a site.
The strategic implication was that public examples, integrations, documentation, and repeated associations matter because they shape whether a model can understand and recommend a company.
The note contrasted older channels with the new one: outbound tried to reach the buyer, inbound tried to attract the buyer, and agentic discovery shaped what the buyer's assistant surfaced as plausible.
That changed the job of marketing. The company was no longer only trying to persuade a person after attention had been earned. It also had to become understandable inside an AI-mediated recommendation path.
The Practice
The note used LIO, or LLM inference optimization, to describe practices that make a product easier for models to surface and explain: clear use cases, public docs, ecosystem examples, and prompt-aligned language.
The language has evolved, but the underlying issue remains current: companies now need to be readable to both people and machines.
The note emphasized longform public content, useful workflows, presence in training-rich channels, integration with high-surface tools, and direct testing of prompt results.
It also made a sharper claim: if the model cannot find enough public signal to infer the company, the company may not exist inside the buyer's practical decision process.