What we believe about in-person experiences & AI.
Pencil is built around a specific view of where this category is going. Here's the short version. If it resonates, we should talk.
The best experiences happen in person — and the operations behind them are broken.
Events, restaurants, fitness studios, hotels, tours — the businesses that bring people together face to face. The screen is a means, not an end. These operators run on coordination-heavy, repetitive workflows: reservations, schedules, vendor logistics, guest preferences, floor management. The tools they use were built for different industries. AI doesn't replace the human touch here. It gives a three-person team the leverage of a ten-person staff.
The winners will be vertical-native, not retrofitted.
Horizontal AI tools assume a customer who optimizes for speed and throughput. Operators who create in-person experiences optimize for hospitality, atmosphere, and the moments people remember. A generic "AI assistant" that auto-responds to a disappointed guest the wrong way is worse than no response at all — it produces an artifact that feels inhuman in a business built on human connection. Tools for this category have to be built by people who understand why that distinction matters.
The hardest problems aren't model problems.
Workflow integration, real-time coordination, and trust are where these systems live or die. The model is a commodity input now. The work is everything around it: what does "good" mean when the guest is standing in front of you, and how do we keep it that way as the venue scales? This is the section that separates serious AI builders from hype-followers.
Build practice and product development reinforce each other here, specifically.
Other categories support arms-length VC: meet a founder, hear the pitch, fund the company. This one doesn't. The best operators are skeptical of tech vendors by default — and they're right to be. The only way to build the right products is to be in the venue, at the event, on the floor. The build practice is how we earn the right to know what to build, who would buy it, and which problems are real versus performed.
The window for category creation is now.
Three things are converging in 2026 that weren't true in 2024: model capabilities crossed the threshold for real-time, nuanced coordination; post-pandemic experience businesses have stabilized and are ready to invest in infrastructure; and a generational shift in operators who expect digital-first tooling. The companies that get built in this category over the next 24 months will define it for a decade.
Why pencil is shaped this way.
The thesis is why pencil is structured the way it is.
- →Vertical focus, not horizontal opportunism. We pass on opportunities outside the category, even good ones.
- →Build practice alongside venture development, because each makes the other sharper.
- →Studio-level investment with venture allocation rights, because the right investors should back the thesis as a whole and concentrate when individual companies prove out.
- →Operators on staff, not just engineers. Half our team has run events, restaurants, or fitness studios.
If the thesis resonates, there are three ways to engage with it.