
Market to AI agents
Consumers are using AI models to help them buy things. First AI models were information aggregators: tools people used to search and summarize. Now they're trusted advisors: what people consult before making a decision. Increasingly, they're becoming trusted buyers: agents that shop, evaluate, and transact on behalf of humans.
When AI makes the ultimate buying decision, then it becomes more important for brands to win over AI than for them to win over the human consumer. In that world, understanding how AI perceives a company's value proposition becomes one of the most consequential things that company can understand about itself.
We apply the principles of AI interpretability research to study how models actually think about companies, and underneath the surface noise we find remarkably stable beliefs, formed from the proof points a company has put into the world (often inadvertently).
Companies starting to take this seriously are realizing something uncomfortable: in an AI-evaluated market, marketing claims without substance to back them up don’t work anymore. Conversely, companies that deliver on their value proposition have unbounded opportunity to reach and acquire new buyers. AI agents are judging every brand's value promise, and are teaching humans how to do the same.
Unusual is the AI interpretability platform for marketing and GTM teams. We help companies shape how AI models actually think about their brand and value proposition. We're backed by Y Combinator and the first investors in SpaceX, Uber, Stripe, Clay, and Notion, and we already work with Big Five agencies and Fortune 100 brands. Demand is outpacing our capacity to serve it.
Our cofounder, Will, is a second-time, venture-backed founder (Khosla and Greylock backed Remedy Health). Will started AI interpretability research in 2014 at MIT, where he applied black-box interpretability techniques to language modeling built on recurrent neural networks. He also built an early prototype of Starlink at SpaceX. Prior to MIT, at sixteen, Will earned international acclaim for having successfully retrofitted his family microwave into a Farnsworth nuclear fusion reactor.
Our cofounder, Keller, was previously Chief of Staff and Head of Growth at 8VC-backed Gatsby. He has made 15,000 cold phone calls as part of the Biden presidential campaign in Pennsylvania; he coauthored, “America, Unite or Die” with statistician and inventor of overnight polling, Douglas Schoen; he studied Econometrics at Princeton, where he was the youngest Div. 1 water polo captain in program history.
Our Head of GTM, Sarah, has scaled two startups from $0 in revenue to eight figures: she drove 10X growth at Circle Medical (YC W17), then built Ambience Healthcare's go-to-market from pre-revenue, which went on to raise $350M from Oak HC/FT, Kleiner Perkins, OpenAI, and a16z. She got her start at Google designing growth engines and strategy and operations at Deloitte.
We've grown fast with a very small team force-multiplied by Claude Code and unlimited token budgets. Our constraint today is time. There are always more high-value things worth doing than the exec team can personally carry: a product surface that needs an owner for a week, a customer relationship that needs attention, a hiring loop that should be run well, a new market worth testing.
This role exists to take any of those, get up to speed fast, and carry it to a successful outcome with little oversight, and to build the systems that make the next version of that work take a fraction of the time. You're the person the executive team relies on to create leverage on the most important initiatives at the business.
You're a high-agency generalist who's at your best when you're handed something ambiguous and important and left to run with it. With little instruction, you take projects from ideation to over the finish line. You are diligent, obsessed with the details, and you can identify what tasks are worthwhile without being told. You are a team player, and you’re eager to do whatever job is required to give your team leverage. You can hold your own in a product discussion, a sales call, and a hiring debrief in the same afternoon, and your reaction is to create structure where there wasn't any. You learn quickly, both by listening and by doing.
Ideally, you're already using tools like Claude Code to do your work better, and you want a role where building the systems that multiply you is part of the work itself. You have good judgment, you can be trusted with sensitive context.
You might have been the most capable employee on every team you’ve ever worked on. That will not be true at Unusual. We’re looking for people that are hungry to work with the best possible people.
Whatever is most important and least owned. It shifts month to month. Here are some of the things we could see you taking up immediately:
Overall, your goal is to take things off the founders’ plates.
If you haven't gone deep on Claude Code yet but you're a fast learner who wants to build that fluency, we still want to hear from you.
Our founders both have experience in quant finance, so we think about talent the way an investor thinks about an asset. A young person with high merit and little experience is an undervalued asset, and the best return they can earn is betting on themselves early. So we hire the way we'd invest. We give high-merit people real range and real ownership long before the market would, and we bet on the team over any one idea. If that's the kind of bet you want to make on yourself, this is the seat for it.
Specifically:
AI agents are the newest audience brands need to speak to, but go-to-market teams don't have the tools to understand how they form opinions or make decisions.
Unusual helps brands in the Fortune 100 and scale-ups like Change.org and Astronomer drive more sales and win more market share by shifting how AI models perceive their offerings.
We work closely with every brand to (1) identify the root cause of AI misperception, (2) align on marketing, sales, and product strategies that appeal to agents, and (3) build infrastructure and content that makes their strongest proof more legible to agents.
Our methodology centers on applying black-box interpretability techniques to AI agents like ChatGPT and Claude to understand how they form opinions about a brand and its competitors. Seeing the why behind each opinion gives us the ability to create targeted interventions that change them.
Our team is ex- MIT, Princeton, and SpaceX
Why work at Unusual?
Culture and Values