UGEN is building an autonomous operating system for commercial UGC. It studies what makes short-form content sell, decides what to make, tests it, and learns from every result.
You stay in charge of what gets published and what gets spent.
Research sits in one place, scripts in another, edits and ads somewhere else, and results in a spreadsheet. Every hand-off is done by people, and the reason behind the work gets lost along the way.
UGEN is built to connect it.
One system carries the evidence from research to brief to test to result, so every decision can be traced back to why it was made, and every result can inform the next one.
Copied by handWhy it worked is lostResults arrive weeks later
Research notesSaved videos, a doc
Script draft v4Another doc
Creator briefEmail thread
Edit and cutdownsEditor’s files
Ad set-upAds manager
ResultsA spreadsheet
One line of reasoning, from signal to result
UGC, then UGEN, then UGENERATE.
UGC
UGEN
UGENERATE
UGCENERATE
UGC
UGEN
UGENERATE
Intelligence before generation.
One system that connects market understanding, strategy, creative and performance, and keeps learning, with you approving what goes out.
In development Being built and tested now
Product vision Planned, not yet available
Illustrative demo Fictional example data
One intelligence, many capabilities.
At the center is UGEN’s intelligence. It connects everything the system understands and decides what deserves your attention.
It starts with the market.
UGEN reads what is happening in your category: the content people make, the campaigns competitors run, and what customers actually say.
Only what holds up moves forward.
Before anything shapes a decision, it has to earn trust. Noise and weak signals are set aside early.
Buying intent over popularity.
UGEN looks for what can sell, not just what gets watched. A viral video is not automatically a good opportunity.
Creative work starts from understanding.
Ideas, hooks and scripts are shaped by what the market has shown, so every creative choice has a reason behind it.
You approve what goes out.
Finished work waits for you. Publishing and spending only happen after you approve them.
Every result makes it smarter.
Outcomes and your decisions flow back in, so each new round starts further ahead than the last.
Market signals
What is happening in your category right now: the content, the competition and the customer.
In development
Market Intelligence
Understands what is working in a category, and why it works.
In development
Evidence Integrity
Makes sure decisions rest on what is real, not on noise or assumption.
In development
UGEN Intelligence
Brings every capability together into one line of reasoning, and knows what deserves your attention.
In development
Strategic Reasoning
Finds the opportunities most likely to sell, not just the ones most likely to be watched.
In development
Creative Intelligence
Turns understanding into ideas, briefs and scripts built for your brand. Content creation follows.
In development
Your approval
Nothing is published and no money is spent until you approve it.
Product vision
Performance Insights
Shows what actually happened, in business terms, after your content goes live.
Product vision
Continuous Learning
Every result and every decision you make leaves UGEN better prepared for the next one.
In development
What each capability delivers
Market Intelligence
Understands what is working in a category, and why it works.
In development
Evidence Integrity
Makes sure decisions rest on what is real, not on noise or assumption.
In development
UGEN Intelligence
Brings every capability together into one line of reasoning, and knows what deserves your attention.
In development
Strategic Reasoning
Finds the opportunities most likely to sell, not just the ones most likely to be watched.
In development
Creative Intelligence
Turns understanding into ideas, briefs and scripts built for your brand. Content creation follows.
In development
Your approval
Nothing is published and no money is spent until you approve it.
Product vision
Performance Insights
Shows what actually happened, in business terms, after your content goes live.
Product vision
Continuous Learning
Every result and every decision you make leaves UGEN better prepared for the next one.
In development
Follow one campaign through the system.
From a handful of videos to a tested creative decision, with every step showing its working.
Illustrative demo · Halvo is fictional
It starts with what is already out there.
UGEN looks at the videos, comments and products shaping a category. This example follows Halvo, a fictional bottle brand.
0:24
3 things in my work bag
@carrylight
@carrylightCaptured 12 Oct
The bottle arrives late
0:18
upside down in my bag since 8am
@mira.commutes
@mira.commutesCaptured 14 Oct
Shows the proof before saying a word
0:31
honest review after a month
@ondesk.daily
@ondesk.dailyCaptured 15 Oct
Promises it, never shows it
Product reference
Halvo Carry Bottle
600 ml, insulated steel
Double-seal lid
Fits a laptop-bag side pocket
“Does it actually not leak if it’s on its side all day?”
@jt_walksDoubt
“Every “leak-proof” bottle I’ve owned has ruined a notebook.”
@notes.by.anaDoubt
“ok where do I get this”
@sam.kInterest
Then it understands why.
Not just what a video says, but why it works: what it shows, when, and what it answers for the viewer.
0:18
upside down in my bag since 8am
@mira.commutes
0:00 – 0:04
The openingShows the result before saying a word.Signal
The productUnmistakable and in frame from the first second.Signal
The messageAnswers the doubt the audience keeps raising.Insight
The feelReads like a real person, not an advert.Insight
The questionWould it work as well from a different creator?Open question
Why it works
The proof arrives before the promise, so the viewer never has to take it on trust.
Still to prove
Something else may explain the attention. That is what the test is for.
Not everything said holds up.
Claims are separated from proof. Only what holds up is allowed to shape the creative.
What was said, and what holds up
This is the Halvo Carry Bottle@mira.commutesHolds upSafe to build on.
“Laptop’s completely dry”@mira.commutesHolds upShown on camera.
“It doesn’t leak”@ondesk.dailyNeeds proofSaid, never shown.
“Mine leaked in my bag”A comment about another brandSet asideOne story, not a pattern.
Only what holds up shapes the creative.
Is there a reason to buy?
An opportunity is judged by whether it can sell, not by how many people watched. Engagement alone is never treated as proof of sales.
Opportunity: proof first, for commuters
The audience doubts “leak-proof” claims. One creator answered that doubt with a demonstration in the first second; the others only asserted it.
Why it is promising
People keep asking whether it really works
Some are already asking where to buy
Only one creator shows the proof up front
Verdict
Worth testing
The bet
Showing proof before the promise will hold more of the right viewers.
Understanding becomes a brief.
Each line of the creative brief carries the reason behind it. Select a line in the brief to see where it came from.
Creative brief
Halvo Carry Bottle · v1
Evidence behind “audience and problem”
@jt_walks comment
@notes.by.ana comment
The doubt the audience raises most often.
Evidence behind “opening hook”
@mira.commutes, opening
The only video that shows proof before any claim.
Evidence behind “product demonstration”
“It doesn’t leak”, said but never shown
@jt_walks comment
An unbroken take answers the doubt without asking viewers to trust an edit.
Evidence behind “supporting message”
Product reference
@mira.commutes
Said in the video and true of the product.
Evidence behind “shot sequence”
@mira.commutes
@ondesk.daily
Keeps the order that worked and drops the slow start.
Evidence behind “claim to avoid”
“It doesn’t leak”, said but never shown
If it is not shown, it is not said.
A test that can teach something.
Two versions that differ in one place, so the result says something clear about what works.
Experiment
Showing proof before the promise keeps more viewers watching.
since 8am. still dry.
A · Proof first
0:00–0:02 Bottle upside down in the bag. No voice for two seconds.
this bottle doesn’t leak
B · Claim first
0:00–0:02 Creator to camera: “This bottle doesn’t leak.”
The only differencefirst two seconds
What changes
Only the first two seconds.
What it teaches
Which opening holds attention, and which one sells.
Sample data Viewers who kept watching. Not real results.
A41%
B34%
0%25%50%
Results start the next round.
What was learned is kept, with its limits, and the next round starts from there.
What UGEN learned
For this audience, proof before promise held attention better in this sample.
One creator, one product. Not yet evidence of more sales.
@mira.commutes
Brief v1
Test A/B (sample)
What comes next
Try the same opening with a second creator.
See whether it carries over to a gym bag instead of a work bag.
Connect real sales results once you approve it.
Back to signals, one lesson smarter
Halvo Carry Bottle Demo campaign
0:24
3 things in my work bag
@carrylight
@carrylightCaptured 12 Oct
The bottle arrives late
0:18
upside down in my bag since 8am
@mira.commutes
@mira.commutesCaptured 14 Oct
Shows the proof before saying a word
0:31
honest review after a month
@ondesk.daily
@ondesk.dailyCaptured 15 Oct
Promises it, never shows it
Product reference
Halvo Carry Bottle
600 ml, insulated steel
Double-seal lid
Fits a laptop-bag side pocket
“Does it actually not leak if it’s on its side all day?”
@jt_walksDoubt
“Every “leak-proof” bottle I’ve owned has ruined a notebook.”
@notes.by.anaDoubt
“ok where do I get this”
@sam.kInterest
0:18
upside down in my bag since 8am
@mira.commutes
0:00 – 0:04
The openingShows the result before saying a word.Signal
The productUnmistakable and in frame from the first second.Signal
The messageAnswers the doubt the audience keeps raising.Insight
The feelReads like a real person, not an advert.Insight
The questionWould it work as well from a different creator?Open question
Why it works
The proof arrives before the promise, so the viewer never has to take it on trust.
Still to prove
Something else may explain the attention. That is what the test is for.
What was said, and what holds up
This is the Halvo Carry Bottle@mira.commutesHolds upSafe to build on.
“Laptop’s completely dry”@mira.commutesHolds upShown on camera.
“It doesn’t leak”@ondesk.dailyNeeds proofSaid, never shown.
“Mine leaked in my bag”A comment about another brandSet asideOne story, not a pattern.
Only what holds up shapes the creative.
Opportunity: proof first, for commuters
The audience doubts “leak-proof” claims. One creator answered that doubt with a demonstration in the first second; the others only asserted it.
Why it is promising
People keep asking whether it really works
Some are already asking where to buy
Only one creator shows the proof up front
Verdict
Worth testing
The bet
Showing proof before the promise will hold more of the right viewers.
Creative brief
Halvo Carry Bottle · v1
Evidence behind “audience and problem”
@jt_walks comment
@notes.by.ana comment
The doubt the audience raises most often.
Evidence behind “opening hook”
@mira.commutes, opening
The only video that shows proof before any claim.
Evidence behind “product demonstration”
“It doesn’t leak”, said but never shown
@jt_walks comment
An unbroken take answers the doubt without asking viewers to trust an edit.
Evidence behind “supporting message”
Product reference
@mira.commutes
Said in the video and true of the product.
Evidence behind “shot sequence”
@mira.commutes
@ondesk.daily
Keeps the order that worked and drops the slow start.
Evidence behind “claim to avoid”
“It doesn’t leak”, said but never shown
If it is not shown, it is not said.
Experiment
Showing proof before the promise keeps more viewers watching.
since 8am. still dry.
A · Proof first
0:00–0:02 Bottle upside down in the bag. No voice for two seconds.
this bottle doesn’t leak
B · Claim first
0:00–0:02 Creator to camera: “This bottle doesn’t leak.”
The only differencefirst two seconds
What changes
Only the first two seconds.
What it teaches
Which opening holds attention, and which one sells.
Sample data Viewers who kept watching. Not real results.
A41%
B34%
0%25%50%
What UGEN learned
For this audience, proof before promise held attention better in this sample.
One creator, one product. Not yet evidence of more sales.
@mira.commutes
Brief v1
Test A/B (sample)
What comes next
Try the same opening with a second creator.
See whether it carries over to a gym bag instead of a work bag.
Connect real sales results once you approve it.
Back to signals, one lesson smarter
From brief to finished cut.
What happens after the intelligence decides what to make: the same Halvo brief, directed, generated, assembled, checked and handed to you for approval.
Product visionIllustrative demo
Video generation is part of our product vision and is not available today.
It starts from the approved brief.
The Halvo brief from above: four shots, ten seconds, one claim to avoid, a reason behind every line.
Direction before pixels.
Framing, lens, light, palette and product treatment are fixed for every shot first, so the scenes stay consistent with each other.
Scenes are made shot by shot.
Each shot is made to the brief and the look, one at a time, so the whole cut stays consistent.
Then it is cut together.
Clips, voice, music and captions are assembled to the ten-second structure the brief asked for.
Weak shots are remade, not shipped.
Every shot is reviewed before you see the cut. Anything that falls short is remade.
Then it waits for you.
The finished cut goes to your approval queue. Nothing is published until you approve this exact version.
Shot 1
since 8am.
Passed
0:00–0:02Low angle into the bag. Bottle inverted, laptop edge in frame.
Shot 2
still dry.
Regenerated
0:02–0:04Hands lift the laptop out. One continuous take.
Shot 3
two seals
Passed
0:04–0:08Macro on the lid. Thread, then gasket.
Shot 4Passed
0:08–0:10Bottle back in the bag. Hold for the end card.
0s2s4s6s8s10s
VideoS1S2S3S4
Voice
Music
Captionssince 8am.still dry.two seals
Input
One approved brief.
Production never starts from a prompt alone. It starts from a brief with a reason behind every line.
Look, fixed for all shots
Lens
35 mm, handheld
Light
Morning window, soft
Palette
Canvas, sage, laptop grey
Product
True to the real Halvo bottle
Generation log
Shot 1made
Shot 2made
Shot 3made
Shot 4made
Assembly
Voice
First line at 0:02: “Still dry.”
Music
Quiet bed, ducked under voice
Captions
Burned in, 2.5 s each
Total 0:10 · 9:16 · captions on
Quality review
On brandThe product looks exactly like the product
Looks realShot 2 fell short and was remade
Sounds rightVoice clear, music sits underneath
Says only what is trueNothing promised that is not shown
The intelligence stays behind the scenes. What you see is a calm workspace: ask UGEN, review the insights, shape the creative and approve what goes out.
A proof-first opening for laptop commuters. Doubt about leaking is what this audience raises most, and the only video that shows proof before the promise is @mira.commutes. I would test it against a claim-first cut from the same creator.
Audience comments
@mira.commutes
What holds up
Why not build on the review video?
@ondesk.daily says “doesn’t leak” at 0:19 but never shows it. I don’t build a brief on a claim the video can’t support.
What would change your mind?
If the proof-first cut does no better than the claim-first cut, I drop the angle and remember why, so it is not suggested again without something new.
Ask UGEN
Shot plan · Cut A, proof first
since 8am.
0–2sInverted in a packed bag. No voice.
still dry.
2–4sLaptop out, dry, in one take.
two seals
4–8sLid close-up. Thread, then gasket.
8–10sBack in the bag. End card.
First line
“Eight hours. Still dry.” On hold until the eight hours are shown on camera
Quality checks
On brand and accurate
Nothing promised that is not shown
Disclosure added before publishing
Insights with their source
Finding
Source
Type
The proof comes before the promise
@mira.commutes
Signal
Doubt about leaking is what buyers raise most
Comments, 14 Oct
Signal
The opening answers that doubt early
UGEN
Insight
“Doesn’t leak” is said but never shown
@ondesk.daily
Signal
Would another creator get the same result?
To test
Open question
Sample data Share of viewers still watching, by second
Contribution after costs
Orders
—
Refunds
—
Production cost
—
Ad spend
—
Contribution
—
Fills in only when a sales connection is approved. Views alone are never reported as success.
since 8am.
Cut A · proof first · v3
0:10 · reviewed and ready
Waiting for your approval
Test budget proposal
Sample: $300 over 5 days, split evenly A/B
Waiting for your approval
You approve the exact version and amount. If either changes, it comes back to you. Demo only: nothing here is published or spent.
UGEN
Cut A v3 for Halvo is ready for you. Reply YES to approve this exact version, or LOOK to review it first.
LOOK
Here it is. The first line is on hold until the eight hours are shown on camera.
Planned channels
Get the daily brief by text, approve exact items from your phone, or ask UGEN out loud: “What happened with the Halvo test?”
Product vision · not yet available
Results sharpen the next decision.
Research shapes the brief. The test produces evidence. The evidence becomes memory, and memory shapes the next round of research. UGEN learns from what actually sold, not from what was merely watched.
Contribution after costs, not views
Your approvals and rejections, with the reasons
Evidence that held up, and evidence that did not
DiscoverFinds the doubt holding buyers back
UnderstandSeparates proof from promise
DecidePicks a proof-first opening
CreateBrief v1, cuts A and B
ApproveYou approve the final cut
MeasureSample: A held more viewers
MemoryEach step leaves a recordThe next round starts smarter
Halvo demo workspaceProduct vision · illustrative preview