The problem
Travel platforms optimize for "lowest price" and "highest rating." Those are useful filters until you're a Black traveler trying to figure out whether the boutique hotel with the 4.9 average and the all-white reviewer base is going to feel like a vacation or a job interview.
For decades, the workaround was social: ask the group chat, follow the right accounts, decode the comment sections. It worked, but it didn't scale, and it required you to do the research labor on every trip. Meanwhile, every major OTA had perfectly competent personalization, for the average user. The average user wasn't us.
What I built
ROAM is an AI travel agent that defaults to cultural fit as a first-class signal, alongside price, dates, and itinerary logic. You tell it where you want to go, what kind of trip you want, and any constraints (kids, accessibility, religious considerations, budget). It comes back with a stitched-together plan, flights, stays, restaurants, activities, weighted by signals that matter to us.
Under the hood, ROAM runs an opinionated retrieval layer on top of a curated index of Black-owned, Black-friendly, and culturally-affirming venues across the U.S. and select international destinations. That index is the actual product. The chat interface is just the front door.
How it works
- Cultural-fit retrieval. Every hotel, restaurant, and experience in the index carries metadata: Black-owned, BIPOC-owned, certified inclusive, community-rated, hair-friendly amenities, dietary fit. The model uses those signals as constraints, not afterthoughts.
- Itinerary generation in the user's voice. The output reads like a friend who knows the city, not a press release. Tone matters, "this spot pops on Sundays" lands differently than "popular with locals on weekend evenings."
- Booking that works. Where we have direct partnerships, you book in-app. Where we don't, we route to the property with one tap.
- Feedback that compounds. Every trip a traveler completes feeds back into the index. The recommendations get sharper because the community is sharpening them.
Why this needed to exist
The "general purpose" travel AI being built by the big labs and the big OTAs is going to scale to the lowest common denominator, because that's where the volume is. None of them will treat "will I feel comfortable here as a Black traveler" as a hard constraint. They might add a tag. They won't reshape the ranking.
ROAM is what happens when a Black engineer with a travel itch and an AI stack decides he doesn't want to wait for that tag.
The product everyone else was going to build eventually, built for us first, on our terms, on our timeline.
Where it is now
ROAM is live at roamblk.com. Travelers are using it to plan domestic and international trips. The index is growing weekly. The team is small, the burn is small, and the product is shipping.
What's next
Three things: tightening the booking funnel where the dropoff is highest, expanding international coverage (the Caribbean and West Africa are next), and rolling out group-trip planning, because traveling alone is one product, traveling with the family or the crew is a very different product.