The Orlando Local
Brought the persona to life with personal insights and local knowledge. Her voice carried the whole experience. Every recommendation came from her, as if she'd been there herself.
Visit Orlando x Tripadvisor · Case study · Conversation design
Designing an award-winning voice experience that made people trust a city they thought they already knew.
★ 2022 Mercury Award for Innovation, U.S. Travel Association
100k
User sessions
Across the US and Canada
3min
Average session length
Deep exploration, not a quick bounce
4.3
Interactions per session
Users stayed and kept exploring
Tripadvisor's first interactive immersive experience in North America. Expanded to Canada following US success. Then to Fire TV. Read the Visit Orlando press release.
At a glance
The architecture decision
Tripadvisor had no template for immersive audio in North America. The safe build was a single narrator reading polished copy. Instead, I designed a dual-voice architecture — a local guide plus a separate voice for verified traveller reviews — running on one reusable loop. Two distinct registers, one consistent structure.
The risk was that two voices can confuse as easily as convince. The fix was a strict separation of roles and a reusable loop that never let either voice bleed into the other.
It returned 100,000 sessions, three minutes average, and became the template for 10+ destination experiences.
The brief
For Visit Orlando, that meant leaving money on the table — travellers who might stay longer, spend more, or return, if they knew the city existed beyond the parks.
Tripadvisor partnered with Visit Orlando to build their first immersive audio experience in North America. The brief was to redefine the city for travellers. Not replace the theme parks. Just show people there was more. Local culture, food, neighbourhoods, the Orlando that people who actually live there know.
A standard tourist guide voice wouldn't do it. It had to feel like a recommendation from someone who genuinely loved the city. Someone you'd trust. It refused to be a tourist-board voicebot.
The persona
The entire experience was built around a single strategic idea: an Orlando local who is passionate about showing off her city. Warm, knowledgeable, proud. Someone who actually lives there and wants you to see the city she loves.
To make her feel real, we cast an actual voice actor from Orlando. Not a synthetic voice, not a generic American accent. Someone who sounded like she'd lived there her whole life. Because she had.
“I love theme parks, but as someone who was born and raised here, I know lots more reasons why you should come for a visit.”
The local guide persona, opening line
Brought the persona to life with personal insights and local knowledge. Her voice carried the whole experience. Every recommendation came from her, as if she'd been there herself.
A second, distinct voice for verified traveller reviews. Real opinions from real visitors, woven in to build trust from outside the brand. Two voices, one experience.
The conversation design
I designed a deterministic, multi-path conversation architecture. Three pathways from a main menu, each opening a sub-menu of three options. Every selection triggers the same loop: local recommendation → optional Tripadvisor review → return to sub-menu or main menu. Nine points of interest, one reusable loop. Built to scale.
PATH A
Italian / Sushi / Gastropub
PATH B
Waldorf / Grand Bohemian / Alfond Inn
PATH C
Feast for the Senses / Non-Stop Thrills / Great Vibes
What changed between drafts
During QA, I caught a continuity error in the shared template: after finishing the sightseeing menu, the skill looped back to the hotels prompt. Fixed by isolating each pathway's exit logic.
“Would you like to explore more of the amazing resort hotels in the area? Just say which between The Waldorf Astoria Orlando, The Grand Bohemian Hotel or The Alfond Inn?”
Draft
“If you'd like to explore some more sightseeing spots, just tell me what you're in the mood for. Just say ‘a feast for the senses’, ‘non-stop thrills’ or ‘great vibes’.”
Final
A pacing checkpoint — an explicit “Ready?” before the main menu — was added to give the user control rather than forcing them straight from intro into navigation.
The Fire TV adaptation wasn't a copy-paste of the Alexa script. Voice confirmations became on-screen buttons. A QR code replaced audio-only detail. The conversation architecture stayed the same; only the modality changed.
User invokes the skill
Soundscape plays. Theme park ambience sets the scene.
“Alexa, launch Visit Orlando.”
The local guide introduces herself
The Orlando Local sets the tone. Warm, local, personal, from the first line. “Ready?” User says yes.
Three pathways presented
Food, Hotels, or Sightseeing. User chooses a category.
Each path opens three options
Food: Italian, Sushi, Gastropub. Hotels: Waldorf, Grand Bohemian, Alfond Inn. Sightseeing: Feast for the Senses, Non-Stop Thrills, Great Vibes.
The Orlando Local speaks
A personal, local recommendation. Specific details. Her own experience woven in. Never generic.
“I know this because I have been going to Prato for years.”
Social proof offered
The Orlando Local asks if the user wants a Tripadvisor review. If yes, Social Proof reads the review. If no, skip ahead.
User chooses what comes next
Return to sub-menu for more in the same category, or say “Main Menu” to start again. The loop runs as many times as the user wants.
Global commands: “Main Menu” returns to the start from anywhere. “Stop” exits cleanly from anywhere.
“I know this because I have been going to Prato for years.”
The Orlando Local, mid-recommendation. Her own life, woven in. Never a brochure.
Verified traveller review
The second voice. Real opinions from real visitors, read in a distinct voice so users always knew who was speaking: the local, or the crowd that agreed with her.
The collaboration
This was not a solo project. It involved developers, sound engineers, and visual designers. My job was to hold the conversation design together across all of them.
Implemented the conversational logic and smooth navigation between voice and screen.
Developed the dual-voice system and sourced authentic ambient sounds for each location.
Provided art direction for the Fire TV multi-modal elements, ensuring audio and visual worked together.
The ripple effect
10+ destination experiences followed, all using the same loop structure I designed. Three examples below.
01
An audio journey through the city's musical roots and live music scene.
02
A personality-driven quiz to match travellers with their perfect island.
03
A cultural showcase built for Amazon Prime, taking the format into a new market.
The experience, running
What this shows
100,000 sessions. Three minutes average. Users who came for theme parks stayed to discover a city. That's what deterministic conversation architecture, a clear voice strategy, and a reusable loop can do at scale.