XENIA

The agents

Two agents at work.
Always on watch.

XENIA is a multi-agent operating system. Echo looks back at the reviews — the voice of the guest. Horizon looks forward at the bookings — the future of the property. Both file into a single briefing, every morning. The agents do the watching. You make the calls.

Live on 1 group · 1 hotel · 412 reviews kept watch

Agent · 01

Echo.

Looks back. The voice of the guest.

Named after Ἠχώ — the Greek nymph who carries voices. Echo aggregates every review platform that mentions your hotels, tags topics on fourteen categories, scores reputation on a proprietary index (ORS), and drafts replies in your voice.

That is the surface. The difference is operational: when negatives point to a real fault on the property — spa closed without notice, AC noise on the same room for the fourth week — Echo flags the pattern, opens follow-up tasks, and keeps them on your radar until the issue is resolved and the complaint stops repeating. A weekly digest lands Friday morning; critical signals move sooner.

Cadence
Weekly digest
Friday 07:00 + ad-hoc alerts
Platforms watched
Booking · Google ·
TripAdvisor · Expedia

01

Reviews aggregation

Booking, Google, TripAdvisor, Expedia. Every review pulled overnight, deduplicated, scored, and stored. No platform-by-platform logging in.

02

Topic detection

Fourteen categories — spa, breakfast, room, staff, noise, value, location, cleanliness, AC, pool, food, bathroom, bed, parking. Each review tagged with its true subject, not just sentiment.

03

Reply assist in your voice

Claude drafts a reply for every review, calibrated on your past responses. You approve, edit, or rewrite in one click. The voice stays consistent across your properties.

04

Problems tracked to closure

Negatives often name real faults — not just mood. Echo surfaces the pattern, opens follow-up tasks, and keeps them visible until the issue is fixed — so the next guest does not repeat the same complaint.

One sample — guest post, your reply, problem tracked to closure

01 · Guest posted on Booking.com

2 / 10Booking.com · Sunday 18:22 · Room 204

Could not sleep — AC noise

Location and breakfast were fine. The problem was the air conditioning in room 204: loud rattle from midnight onward. Reported it twice at reception during our two-night stay — nothing changed before checkout.

Topics tagged: Room · AC · Noise · Reception

02 · Your reply — drafted overnight, in your voice

Apology first. No internal ops on the public thread — you copy and paste on Booking when ready.

Dear guest, thank you for taking the time to share your experience. We are sincerely sorry the air conditioning in room 204 disrupted your sleep — that is not acceptable, and we are sorry reception did not resolve it during your stay. Your feedback matters to us. We hope to welcome you back under better conditions. Kind regards, The Management

Approve & copy to Booking →or edit · or rewrite · save to your record

03 · Problem tracked inside XENIA — A to closure

For your team only: pattern flagged, suggested actions, status until the fault stops repeating.

Persistent AC noise in room 204 — reported twice at reception during a two-night stay, no fix before checkout. Fourth negative mention on the same room this month.

Suggested actions

  • Schedule HVAC inspection on unit 204 before the next arrival
  • Brief night reception: escalate noise complaints to the duty manager within 15 minutes
  • Mark the problem done once the vendor confirms the unit is quiet
ValidatedIn progress→ Done when the next guest on 204 does not mention AC

Agent · 02

Horizon.

Looks forward. The future of bookings.

Named after ὁρίζων — the Greek boundary between what was and what's coming. Horizon reads your bookings overnight, projects landing occupancy, watches your comp-set on Booking.com, and names the pricing windows that matter — before they close.

Cadence
Daily briefing
08:00 + comp-set refresh
PMS connectors
Mews · Cloudbeds · OfiHotel · Opera
Plus Apaleo and Protel · read-only

01

Occupancy MtD + landing

Read bookings overnight from your PMS. Project landing occupancy for the current month, named by hotel and segment — not buried in a 12-tab spreadsheet.

02

Comp-set on Booking.com

Daily comp-set scrape on Booking. Per night, per room type, per competitor. Soft-spot detection : when a competitor is selling and you are not, you hear about it before noon.

03

Morning briefing 08:00

One paragraph. 90 words. The week's pressure points, today's pricing window, the decision needed before lunch. Read with coffee. No dashboard required.

Forecast · next 7 days

82

Mon

85

Tue

76

Wed

58

Thu

61

Fri

79

Sat

88

Sun

Each square = projected occupancy %. Orange = soft spot, olive = healthy. The agent watches the gap, the briefing names the action.

Morning briefing · Monday 08:00 · sample

“Occupancy MtD 71% (target 75). Thursday–Friday next week trending soft : 58% / 61% projected, vs comp-set selling at 78%. Hotel Vincci dropped €15 on the same room category yesterday — your rate is now €22 above the median. Pricing review recommended before Wednesday noon, or accept the gap. Echo flagged a 2-star review on spa noise (Room 204) overnight — drafted reply waiting in your inbox.”

Generated by Horizon · validated by Echo · delivered as email + Slack

How the two agents pair

Past and future. Voice and vision.

Each agent has its own angle on the same property. Echo turns the past 7 days of reviews into one Friday digest. Horizon turns tonight's PMS data into tomorrow's morning call. The pairing is the product — neither agent ships alone.

Echo

The voice of the guest

Horizon

The future of bookings

Regard
Past
Future
Signal
The voice of the guest
Booking data
Cadence
Weekly digest
Daily briefing
Buyer persona
Marketing · Brand
Revenue · Commercial
Output
ORS · topics · drafted reply
Occupancy · forecast · comp-set

Talk to us

Tell us about your group.

Bring us your number of properties, your PMS, your weekly rhythm. We come back within 48h with a pilot plan tailored to your operation.

Book a demo →
Direct line · founderhermes@khelys.dev