Personal Agents & Monetization

Lucas UrbisaiaOct 12, 2026

Below is a more extensive writeup on my tweet about personal AI and monetization: HOT TAKE: the only way to monetize personal AI agents will be via transaction fees. By process of elimination: Subscriptions are a no-go because they introduce friction to adoption. If you’re selling personal AI, then you want to lower the barrier to getting users. Users are doubly important: they drive both revenue and data. Data is crucial to making a best-in-class AI product… So you need many users, a lot of data, and borderline-negative friction to get as many of them as possible. That leaves ads and transactions. The problem with ads is that the value proposition of a personal agent comes from a user trusting that the agent is THEIR representative. Ads are in the interest of the platform/advertiser. When recommendations are influenced by advertisers, consumers become skeptical of search results. Google solved this by showing 10 organic results per page, and 1-4 ads at the top. A personal agent serves you 1 to 5 options through a chat interface… space is more limited, and attention gets routed to fewer results. It’s unlikely that users will re-ask the agent for more results. How many times would you re-ask your assistant for new results until you get tired of waiting? Chances are you will open a google search by then. So users need to trust that the results a personal agent displays are unbiased. That only leaves room for transactions. The intriguing part about transactions is that they bring up a two-pronged challenge. Agents may be capable of searching the web, but what infrastructure exists to show them the best results for their searches? None yet. I predict that a commerce network built to aggregate merchants will emerge as agent transactions grow. This network will funnel agentic expenditures through a single layer. This is the market’s best solution: it maximizes transaction volume (good for the platform), allows consumer AI to monetize (good for whomever builds on the consumer side), and aligns incentives for both consumers and businesses. For consumers, the network provides agents with hyper-personalized searches. It borrows merchant historical data (like location, reviews) and real-time data (inventory, open/closed) to provide agents with the data they need to offer optimal search results. In the aggregate, the network also allows consumers’ agents to negotiate discounted pricing and merchant loyalty programs. Merchants benefit too. Realistically, merchants cannot trust a company like Amazon or Shopify to display their results in a world where they compete for 1-5 spots on a texting interface. Everyone is competing on less available real-estate than on a dedicated shopping screen. The outcome is that relevance and personalization win. For good or for ill, those two require some type of gateway that can tap into a company’s own data to expose information as it is needed. Think about inventory and courier availability, for example. That's real-time data that needs to be provided on the business side to ensure a consumer agent will prioritize that search result (recall that the consumer agent purpotedly represents the individual consumer). Where does that leave us? On one hand, merchants and companies that can collect operational data and expose it autonomously will gain an advantage. Though I see that advantage eroding and, eventually, becoming a staple of businesses' technology stack. On the other hand, some type of gateway technology company will emerge. Whether it will partner with existing platforms, like Amazon, or figure out logistics, remains unclear to me. But I do see this as necessary to allow businesses to interact with agents at scale. The outcome, if successful, is that transactions will drive the brunt of revenue for the consumer-agent economy.