Localizing an AI product for American users.

[One sentence: what made the U.S. market hard for this product?]

[2–3 sentences of context: what Fuzozo is, who it was originally built for, and why its experience did not translate directly to American users.]

Five U.S. personas, five distinct audiences.

I led UX research to understand what U.S. users actually needed, then built five personas that defined distinct target audiences and guided localization priorities.

[Methods: interviews, surveys, competitive review… and how many people you spoke to.]

• Persona 01 [Name / archetype]
• Persona 02 [Name / archetype]
• Persona 03 [Name / archetype]
• Persona 04 [Name / archetype]
• Persona 05 [Name / archetype]

From findings to product choices.

Each research finding was turned into a concrete product or interaction decision.

1. [Decision 1: what changed in the product or interaction] [The research insight that led to it]
2. [Decision 2: what changed in the product or interaction] [The research insight that led to it]
3. [Decision 3: what changed in the product or interaction] [The research insight that led to it]

Teaching the AI to sound local.

I tuned the AI’s prompts so its responses matched U.S. user expectations and localization requirements.

Before	After
[An example response before tuning, anonymized]	[The same moment after tuning, and what changed]

[The result, in one sentence.]

[What shipped or changed, any feedback or metrics you are allowed to share.]

Featured in Mashable’s CES 2026 coverage ↗

[What you learned about localizing AI.]

[2–3 sentences: what you would do differently, and what this taught you about designing AI for a new culture.]