Connecting the dots to see and build what's next.
Figment reads the world's signals and finds the opportunities worth building. The Foundry turns them into products — and the best into companies. One loop, operator-funded, approaching full autonomy.
Ideas in. Companies out.
Ideation engine
- Detects weak signals before they're obvious.
- Tracks how trends travel country to country and predicts the next stop.
- Scores every idea against what the studio can already build.
Build engine
Reading the whole world, every day
A topic that shows up across several sources at once counts as a stronger signal than a search spike alone.
26 feeds read daily · 8 more ready to switch on. Hover or tap a category.
Where trends travel.
See a trend in Tokyo before it lands in San Francisco. Figment follows each trend across 65 countries, learns the usual routes and the days between stops, and predicts where it goes next.
A hop is a new week the trend broke out somewhere new (interest 50+ out of 100).
Scored against tech we already own.
Growth that holds week over week, in more than one country.
Each product line pitches ideas that fit what it already does.
Every idea is scored on how much of our platform it reuses.
A reviewer flags the mediocre ones, so only the strongest survive.
From “insurance confusion” to PoliSays.
Insurance stayed one of the steadiest searches of 2026: US searches for car insurance quotes rose about a quarter from winter, and “insurance denial” peaked in May. PoliSays reads home and auto policies section by section in plain language, quoting every clause.
- 01SignalAll of 2026
- 02PeakSpring 2026
- 03BuiltSame platform
- 04LiveSummer 2026
Every launch makes the next pick smarter.
Daily scans across search, news, papers, filings, launches and social — in 65 countries.
Scout agents pitch ideas per product line; a skeptic agent kills the mediocre ones.
Figment writes a buildable blueprint mapped to tech the studio already owns.
The Foundry assembles product, ops, marketing and support from shared agents.
Live on a real domain in weeks, measured against the thesis that started it.
Echo feeds real results back into Figment's reviewer, so every launch sharpens the next pick.
Your users talk. Your product changes. Nobody files a ticket.
Once a company is live, its users keep shaping it. In a client's implementation of our Agent Studio, a user records a thought. A relay of our agents takes it from voice note to a live update, with no tickets and no hand-offs.
- -Feedback sits in a form, a chat or an inbox
- -Someone retypes it into a ticket
- -It waits for a product meeting and a sprint
- -Engineering, QA and security each take a turn
- -The user who asked never hears back
- The user's own voice is the starting point
- Agents sort it, plan it and challenge it
- Building, QA and a security review are part of the run
- The user gets a reply in the app when it ships
- Every note teaches the next run, so the loop gets sharper
Most tools cover one piece: feedback, tickets or code. This runs from the spoken note to the live change.
The person who asked hears back in the app, and every exchange is kept.
Results flow back to Athena, our learning engine, which shapes future plans.
Longer if the next sprint is already full.
Built, tested and deployed.
"The save button is hard to find on my phone. Could it stay at the bottom?"
"Done. The save button now stays at the bottom of the screen. Thanks for the note."
- 01Talk
A user records a voice note about the product, right inside the app.
- 02Hear
The note is transcribed and sorted into bug, new feature or usability.
Agent Studio · Transcribe & sort - 03Plan
Everest, trained on many hours of product and UX teaching, shapes the request into a spec. It lands on the task board with test steps.
Everest · Product & UX leadStudio agent · Board & updates - 04Challenge
A skeptic pushes back before anything is built: is this the right fix?
Sentinel · The Gate - 05Build
The change is designed and built, then checked against the spec.
Build agent · CodeStudio agent · QAEverest · UX review - 06Secure
Every change is scanned for security issues before it ships.
Aegis · Cybersecurity - 07Ship & tell
It goes live, and the user gets a plain-language note in the app saying it's done.
Hermes · Release notesArgus · The Watcher - 08Learn
What worked and what broke becomes a rule, so the next request goes faster.
Athena · Learning engine
The flow is shown as designed. Live counts of requests and time to ship will appear here once we track them.
Incubators pick founders. This picks markets — then builds.
A thesis that survives scrutiny gets built the same week.
Each company starts on tooling the studio already runs in production.
Outcomes train the reviewer; the hit rate is designed to rise over time.
Others find ideas or write plans. We ship companies.
AI venture studios are now a real category. Here's what only this loop does.
Most studio tools stop at a plan or a pitch. We take ideas to live products and registered companies.
Waves follows one trend country to country and predicts where it lands next. We haven't seen anyone else do this.
CallOrPut, PetConnect and AlienSquad are registered companies, PoliSays is forming, and every studio product is live on a real domain.
Ideas are ranked by how much of our own platform they reuse, not by generic market size.
First Seen logs when each signal appeared, so results can be checked against the date we spotted it.
Not a single US source: search, news, research, launches and social across 65 countries.
