Everyone will have a personal agent

Short thesis on why personal agents are inevitable and how they will create a $1T market

October 7, 2026 · 3 min read · Also published on 𝕏.

Personal agents bring consumer friction to 0

Text is the new interface for doing anything high-intent on the internet. A regular consumer will never spend any time researching what the best home appliance or doctor is, or how to file a form with their local government office.

Inbound chores will also be a thing of the past soon. Any ticket, bill or legal request will be answered for you by your agent. This is the first time when I’m not behind on paying my Wi-Fi bill (that doesn’t support recurring payments) because I just have my agent doing it.

Things will also get so much cheaper for a regular consumer. An agent will cancel every unnecessary subscription, chase a refund and negotiate a price for the big purchase you are about to make. I saved over $3000 this way over the past 2 weeks, after I sent my agent to negotiate 2 SaaS bills that we accidentally consumed too much of. Agents can do the ugly work that no one wants or has time to do.

New $1T market

Businesses that don’t support personal agents will soon look like businesses without a website. For a business that doesn’t accept agents, maintaining a relationship with its customer will be harder than ever. Just like with ecommerce, the customer will go to the place that is the most accessible and the cheapest. It was a website before; it will be an agent now.

This will cause a massive reshuffling and a lot of demand for new tools, distribution, marketplaces and ways for the merchants to keep the relationship with their customers. Just like ecommerce created demand for Meta ads, Stripe and Shopify, agentic commerce will create new winners that will help businesses to find their place in the new economy.

Winners are not decided at all

On the consumer front, we have incumbents like Meta and OpenAI and startups like Instinct (or Boski, but we are not really in the conversation right now). Meta has capital, distribution, existing social graphs and Mark Zuckerberg. OpenAI has compute and a great number of people already using ChatGPT. Startups have no previous product debt or existing customers they need to deliver for and get to iterate faster and be closer to the trends. The total market penetration is still tiny; we’ve only had those agents for a ~month.

On the business side, there are active tensions between existing apps and their relationship to agents. Some (Amazon, Booking) are scared of agents and are actively blocking them. Their competitors (Shopify, Expedia) are natively integrating with personal agents. On all fronts, there will be a need for new tooling, services and tools to grow their business in the new reality.

Where is the money

Something I didn’t predict at the beginning was the amount of inference that personal agent companies are able to subsidize so early on. It’s basically impossible to make money in this market reselling inference or the experience of using an agent.

Consumer inference is not only a low-margin business; it’s an extremely negative-margin business that will help drive the advertising and fintech parts of the business. And as the personal agent companies lose money serving their customers, there will be a whole new economy of plugins and integrations that help the consumers and merchants find each other in the new life. And the plugins, marketplaces and B2B tooling will have the good old software margins.

The future is exciting

This feels like a new reshuffling of the internet like Web 2.0 and mobile. I was wrong about this once (I used to deeply believe Farcaster would become the most important social network in the world), but this feels like an incredible opportunity to build a consumer company that will have an impact on the global economy. Boski will play a big part.