Speak the market’s language
Match U.S. search behavior and visual expectations without flattening Combate’s voice.
01 / At a glance
A U.S. growth strategy turned into a content system the team could run.
- Client
- Combate Global, a bilingual fight-sports media brand
- Challenge
- Expand U.S. discovery without weakening the existing Hispanic audience
- My role
- AI workflow, content strategy, analytics, and growth operations lead
- What I built
- Custom AI agent, bilingual content system, QA workflow, and performance dashboard
- Starting asset
- 1,888 videos, 172.2M lifetime views, and 413K subscribers
- Delivery
- A working AI-assisted operating system launched in 30 days
02 / The problem
How do you reach a new U.S. audience without losing the audience you already have?
Combate had a powerful media asset: a deep catalog of fighters, fights, and storylines. But most of the library was packaged for Spanish-speaking audiences across Hispanic markets.
Simply translating the archive into English risked stripping away the identity and context that made the content valuable in the first place.
The challenge was not translation. It was expansion without erasure.
03 / Strategic choice
Treat every video as a market, creative, and data decision.
The solution had to connect market intelligence, AI-assisted production, bilingual judgment, and performance measurement in one operating system.
Protect what made it valuable
Preserve fighter identity, Latin American storylines, Spanish context, and the community relationship.
Optimize the complete package
Make titles, descriptions, timestamps, and cover direction communicate value to viewers and YouTube.
Build a system, not a batch
Make recommendations responsive to performance, geography, search demand, competitors, and market signals.
04 / What I owned
From market signals to a system the content team could operate.
Shape
The growth strategy
Mapped U.S. demand using geography, search behavior, traffic sources, retention, competitors, and fighter interest; then defined the English-forward bilingual strategy.
Build
The AI workflow
Built a custom Combate agent connected to YouTube Analytics and external market data, with outputs for titles, descriptions, timestamps, and cover direction.
Operate
The learning loop
Created human-review rules for accuracy, bilingual tone, brand voice, and cultural context, then connected publishing results to the next optimization cycle.
05 / The system I built
Six connected layers turned raw audience signals into bilingual content decisions.
Content context, audience data, market demand, creative production, and publishing results worked as one continuous loop.
01
Library context
Content archiveUnderstand what each video contains. Video metadata, fighters, countries, storylines, and key fight moments established the context behind every asset.
02
Audience performance
YouTube AnalyticsSee how viewers already behave. Views, watch time, retention, traffic sources, and audience geography revealed existing strengths and missed opportunities.
03
Market intelligence
Search · competitors · fighter demandFind demand the channel was missing. U.S. search behavior, competitor packaging, and fighter interest exposed new paths into mainstream fight-sports discovery.
04
AI analysis
Custom Combate agentTurn multiple signals into one recommendation. The domain-specific agent synthesized content, channel, and market data through Combate’s bilingual brand strategy.
05
Content production
Packaging systemOptimize the complete discovery package. The system produced English-forward titles, bilingual descriptions, key timestamps, and cover-image direction.
06
Review and learning
Human QA · dashboardProtect the brand and improve the next cycle. Human review checked fighter accuracy, tone, and cultural context. The dashboard fed U.S. and Hispanic audience response back into the system.
06 / Results
U.S. views grew 55.5%—faster than the channel overall.
The signal
The channel did not trade one audience for another. U.S. view share increased by 3.8 percentage points while total views, watch time, and engagements all grew by more than 30%.
239K U.S. views in the April window—up from 153K in the matched March window.
| Metric | March 1–27 | April 1–27 | Change |
|---|---|---|---|
| U.S. views | 153K | 239K | +55.5% |
| U.S. view share | 21.2% | 24.9% | +3.8 pp |
| Total views | 725K | 957K | +32.0% |
| Watch hours | 119.5K | 157.6K | +31.9% |
| Engagements | 7.68K | 10.27K | +33.8% |
| Average view duration | 10:20 | 10:51 | +31 sec |
Measurement note: matched 27-day windows, March 1–27 versus April 1–27. Results reflect the combined content program, not a controlled single-variable test.
07 / What this proves
Strategy and execution stayed connected all the way through audience response.
AI Tooling & PrototypingBuilt a domain-specific agent inside a live content workflow.
Creative ProductionCreated bilingual packaging across copy, timestamps, and cover direction.
Technical Project ManagementConnected data, AI, editorial QA, publishing, and reporting.
Go-to-Market OperationsTurned a U.S. expansion strategy into measurable content operations.
Zero-to-OneBuilt a working AI-assisted operating system in 30 days.
Community & EcosystemExpanded discovery without abandoning the existing fight community.
Archive1,888 videosA system designed for scale, not one-off copywriting
Audience base413K subscribersAn established community whose trust had to be protected
Lifetime reach172.2M viewsA valuable catalog reorganized for new discovery
08 / Takeaway
The operating system was the strategy made real.
The hard part was not generating English copy. It was expanding U.S. discovery without stripping away the context that made the archive valuable.
I turned that tension into a working system—connecting market strategy, AI tooling, creative judgment, operational delivery, and measurable audience growth.