How Startups Can Build Brands for Both AI Agents and Human Customers

Brands are no longer writing only for humans. Toby Coulthard describes a 2026 shopping scenario in which a person shares a room photo, budget and style with an AI agent that can recommend furniture and even purchase it on the person’s behalf.

The key takeaway: AI agents may evaluate product pages before a human ever sees them. That means brands need messaging that remains clear and useful when interpreted, summarized or reassembled by a machine.

Coulthard argues that startups should define a distinct voice and write it down, rather than leaving it only in a founder’s head. This should include not just what the brand sounds like, but also words it should never use and claims it should never make.

That documentation matters because AI tools follow explicit rules and can drift without them. Newer brands have an advantage because they can build agent-ready copy and guidelines from the start instead of retrofitting older systems later.

Product descriptions are no longer just administrative copy; Coulthard frames them as core brand assets in an agent-mediated world. If agents read descriptions to recommend or purchase products, the source description needs to preserve what makes the brand distinct.

Marketing teams should also check how AI summarizes their messages. Gmail and Apple are cited as current examples where summaries may appear before the recipient opens the email, so teams should verify whether the offer and differentiation survive the summary.

As AI makes content easier to produce, Coulthard warns that measuring marketing by volume alone will become less meaningful. Teams should focus on business outcomes, conversions and revenue to understand whether messaging is actually working.

The article also cautions against removing humans from the equation. AI agents still serve user preferences, while humans remain the primary readers and decision-makers today.