86 Labs pulls your systems, your guest requests and your conversations into one place. Your team spends the day with guests instead of chasing what slipped through.
A guest mentions an anniversary to the concierge. The restaurant never hears about it. A maintenance issue surfaces at midnight and waits on a handwritten note until the morning briefing, if it survives that long. A debriefing takes more than thirty minutes because someone has to rebuild it from five systems that have never spoken to each other.
You are not losing to another piece of software. You are losing to WhatsApp groups, Post-its, and a briefing PDF someone assembled by hand during night shift.
73 percent of travelers prefer hotels that personalize their stay using data the guest already gave them. Only 30 percent feel hotels actually deliver it. That 43-point gap is not a mystery. It is a preference that never made it from the concierge to the restaurant.
That is not an edge case. It is every single week, and no amount of staffing solves it on its own.
86 Labs sits on top of the systems you already run, and gives your team back the hours they spend hunting for information. Three modules. That is the whole platform.
Shadow listens to every guest conversation, phone calls, WhatsApp, texts, in any language, and turns what is said into a task for the right department instantly, whether a human took the call or Shadow's own AI voice did.
Active mode — a multilingual AI voice agent answers calls around the clock, logs the request, and creates the task on the spot. No guest waits on hold because of a language barrier.
Shadow mode — listens silently in the background of a human-staffed call and generates the task automatically. Nothing said gets lost because someone forgot to write it down.
Screen-pop — an incoming call is matched to the guest's profile before your team even picks up. They know who is calling, and why it matters, before they say hello.
Isabelle checks in alone, in town for a work meeting. Nobody at the front desk asks about her anniversary, because nobody needs to: last time she stayed, she mentioned to a server, over dinner, that this was the date she and her husband got married.
She is upgraded to a better room to celebrate that special date.
Instead of the usual amenity, there is a card and a pen on the desk: a quiet way to write something to her husband, on the one night of the year that matters to them, even hundreds of miles from home.
She calls. Shadow listens in the background, and a priority ticket reaches engineering before she has even hung up.
Pulse hands the GM the night in one page: one incident, resolved within the hour.
It mentions the letter by name.
Recommendations are always suggestions, not instructions. The staff member decides. No module removes a human role unless the hotel chooses it explicitly.
We document what each module does, what data it uses and why, as part of the signed DPA. We do not claim our outputs are perfect; known limitations are communicated.
In shadow mode the human team stays on the line throughout; Shadow generates tasks and notes and never intervenes unless triggered. Active mode is enabled only by explicit decision of the hotel, with caller wording agreed in advance. We never instruct a system to make a guest believe they are speaking to a human where that confusion would be harmful. Transcriptions are retained for the period the hotel configures, 30 days by default, and voice data is never used to train third-party models without written consent.
We do not sell, share or monetise guest data, and do not cross-reference it between hotels without consent from both parties. At the end of a relationship all data is deleted within 30 days, with written confirmation.
Recommendations are based on behavioural stay data only, never on protected demographic characteristics. Models are reviewed regularly, and any bias raised by a client is investigated within 30 days with a report provided.
Security incidents affecting hotel data are notified within 48 hours. Daily backups with quarterly restoration testing. No major update ships without a prior testing period and advance communication.
This charter is reviewed at least every 12 months. Ethical questions go to mathieu@dnd86.com and are answered within 15 business days.<br>Version 1.0 / last updated August 2026.

86 Labs pulls your systems, your guest requests and your conversations into one place. Your team spends the day with guests instead of chasing what slipped through.
A guest mentions an anniversary to the concierge. The restaurant never hears about it. A maintenance issue surfaces at midnight and waits on a handwritten note until the morning briefing, if it survives that long. A debriefing takes more than thirty minutes because someone has to rebuild it from five systems that have never spoken to each other.
You are not losing to another piece of software. You are losing to WhatsApp groups, Post-its, and a briefing PDF someone assembled by hand during night shift.
73 percent of travelers prefer hotels that personalize their stay using data the guest already gave them. Only 30 percent feel hotels actually deliver it. That 43-point gap is not a mystery. It is a preference that never made it from the concierge to the restaurant.
That is not an edge case. It is every single week, and no amount of staffing solves it on its own.
86 Labs sits on top of the systems you already run, and gives your team back the hours they spend hunting for information. Three modules. That is the whole platform.
Shadow listens to every guest conversation, phone calls, WhatsApp, texts, in any language, and turns what is said into a task for the right department instantly, whether a human took the call or Shadow's own AI voice did.
Active mode — a multilingual AI voice agent answers calls around the clock, logs the request, and creates the task on the spot. No guest waits on hold because of a language barrier.
Shadow mode — listens silently in the background of a human-staffed call and generates the task automatically. Nothing said gets lost because someone forgot to write it down.
Screen-pop — an incoming call is matched to the guest's profile before your team even picks up. They know who is calling, and why it matters, before they say hello.
Isabelle checks in alone, in town for a work meeting. Nobody at the front desk asks about her anniversary, because nobody needs to: last time she stayed, she mentioned to a server, over dinner, that this was the date she and her husband got married.
She is upgraded to a better room to celebrate that special date.
Instead of the usual amenity, there is a card and a pen on the desk: a quiet way to write something to her husband, on the one night of the year that matters to them, even hundreds of miles from home.
She calls. Shadow listens in the background, and a priority ticket reaches engineering before she has even hung up.
Pulse hands the GM the night in one page: one incident, resolved within the hour.
It mentions the letter by name.
Recommendations are always suggestions, not instructions. The staff member decides. No module removes a human role unless the hotel chooses it explicitly.
We document what each module does, what data it uses and why, as part of the signed DPA. We do not claim our outputs are perfect; known limitations are communicated.
In shadow mode the human team stays on the line throughout; Shadow generates tasks and notes and never intervenes unless triggered. Active mode is enabled only by explicit decision of the hotel, with caller wording agreed in advance. We never instruct a system to make a guest believe they are speaking to a human where that confusion would be harmful. Transcriptions are retained for the period the hotel configures, 30 days by default, and voice data is never used to train third-party models without written consent.
We do not sell, share or monetise guest data, and do not cross-reference it between hotels without consent from both parties. At the end of a relationship all data is deleted within 30 days, with written confirmation.
Recommendations are based on behavioural stay data only, never on protected demographic characteristics. Models are reviewed regularly, and any bias raised by a client is investigated within 30 days with a report provided.
Security incidents affecting hotel data are notified within 48 hours. Daily backups with quarterly restoration testing. No major update ships without a prior testing period and advance communication.
This charter is reviewed at least every 12 months. Ethical questions go to mathieu@dnd86.com and are answered within 15 business days.<br>Version 1.0 / last updated August 2026.
