Not Done Weekly - Breaking Down Steps for AI Agents for Hospitality

Harman's throwaway stat — that 56% of a front desk agent's shift is administrative — is actually a framework in disguise. Every role in your hotel is some blend of three buckets, and until you've decomposed the roles, every AI conversation you have is premature. Here's a five-step exercise you can run this week on a single role.

The Three Buckets (memorize these)

1. Administrative / Transactional — data entry, system clicks, reconciliation, templated emails, report pulling, rate loading. Rules-based, repeatable. AI eats this.

2. Interpersonal / Judgment — reading a guest, coaching a teammate, negotiating a group, making a comp call, defusing complaints, overriding the system. AI amplifies this.

3. Physical — stripping rooms, plating food, walking a VIP, fixing a leak, setting a ballroom. AI doesn't touch this YET (key word being “yet”).

Step 1 — Pick the Role (15 minutes)

Don't start with front desk. Everyone's starting there. Pick a role where you genuinely don't know the admin percentage. Good candidates: reservations agent, revenue manager / analyst, group sales coordinator, night auditor, AP / AR clerk, recruiter or HR coordinator, catering sales manager.

Step 2 — Capture a Full Shift (1 shift, 1 day)

Pick one of three methods based on what's realistic: shadow them (sit next to the person for one full shift with a notebook, log what they're doing every 15 minutes), have them self-log (give them a one-page template with 15-minute blocks and ask them to jot one line per block — tell them why: this is about designing their role for the next 3 years, not evaluating them), or pull system logs (for digital-heavy roles like reservations, revenue, accounting, pull the clickstream or system activity for one representative day).

Step 3 — Bucket Every 15-Minute Block

Go through the shift log and tag each block: A = Administrative/Transactional, I = Interpersonal/Judgment, P = Physical, W = Waiting / idle / context switching (track this separately — it's often 10-15%). Add up the percentages. Write them down.

Step 4 — Interrogate the Admin Bucket

For every admin task in Bucket A, ask four questions: Is this task rules-based (same inputs → same outputs every time)? Is this task pattern-repetitive (happens multiple times per shift or week)? Does this task require data that already lives in a system (PMS, CRS, RMS, CRM, email)? Would removing this task reduce guest or owner experience?

If the first three are yes and the fourth is no → that task is AI-addressable today. Circle it. That's your automation backlog for that role.

Step 5 — Answer the Repositioning Question

This is the step almost everyone skips, and it's the one that matters most. For the hours you'd free up, answer: If this role got 30% of its time back, what would I want them doing instead? Does that new activity drive revenue, loyalty, or retention? Can I measure whether they're actually doing it? Does the person in the role have the skills to do the higher-value work — and if not, what training closes the gap?

If you can't answer these clearly, you're not ready to deploy AI against the role. You're ready to have a design conversation first.

The Scoreboard

After you run the exercise, you should be able to fill in this one-liner for the role:

"[Role] currently spends ___% of their time on administrative work. If we remove ___% of that through automation, we'll redirect those hours toward [specific revenue or experience activity], which we'll measure via [specific KPI]."

If you can write that sentence, you have a business case. If you can't, you have more work to do before you sign an AI Agent contract. Most hoteliers evaluate AI vendors role-by-role in a vacuum, don't do that — you'll end up buying the loudest pitch instead of the highest-leverage problem. Decompose three roles first, stack-rank them against each other, and let the data tell you where to spend. 

^{The Hoteliers' AI Vendor Field Guide}

A practical, category-by-category map of who's actually building AI for hotels right now. This isn't an endorsement list — it's a reference.

Guest Engagement & Messaging (chat, SMS, web, in-stay)

The AI layer between your hotel and your guest, before, during, and after the stay.

• Canary Technologies — Guest management system: web chat, AI voice, AI messaging, mobile check-in, mobile keys, F&B ordering, contactless checkout. Live in 20,000+ hotels across 100+ countries. Canary
• Akia — AI guest messaging platform, popular with boutique and independent hotels.
• HiJiffy — Multilingual chatbot and voicebot focused on direct booking conversion and PMS/CRS integration.
• Asksuite — Manages 85% of customer requests, supports 37 languages, connects with 100+ systems. Dialzara
• Conduit — Voice AI + messaging across phone, SMS, WhatsApp, email, and OTA channels in one inbox.
• Visito — Integrates with WhatsApp, Instagram, and Messenger to automate over 97% of guest messages. SiteMinder
• EasyWay — AI Concierge, AI Reservation Manager, AI Receptionist; supports 100+ languages.
• Myma.ai — GPT-powered guest messaging across WhatsApp, web, email, and OTAs with PMS integrations.
• Kipsu — Human-led messaging platform; positioned for luxury where brand voice matters more than scale.
• Revinate Ivy — AI-driven SMS and messaging channel with 98% open rates. Tossom
• EVA (Fourteen IP) — AI-powered virtual assistant; voice + SMS/WhatsApp/Facebook Messenger/webchat.
• Duve — Pre-arrival, in-stay, and post-stay guest journey messaging.
• Whistle (by Cloudbeds) — Guest messaging integrated into the Cloudbeds platform.

Voice AI (phone, reservations, front desk)

Where the labor pressure is highest and the OTA leakage is most visible.

• Canary AI Voice — Front desk and call center inquiries, reservations, FAQs, dining recommendations.
• PolyAI — Enterprise conversational AI for hotel groups managing thousands of calls per day.
• Aiello (AVA) — Voice + concierge; supports 58 languages via GuestWeb; hotels report 43% drop in front desk call volume. Dialzara
• HotelPlanner Booking Assistant — High-volume voice AI handling 10,000+ daily interactions across 15 languages.
• Cloudbeds Engage — Voice + SMS automation natively integrated into the Cloudbeds ecosystem.
• The Hotels Network — KITT — Conversational voice + text guest service agent with brand voice customization.
• Revinate Reservation Sales — AI-assisted voice sales tools achieving 4x average stay value compared to OTA bookings. Tossom
• roommaster Concierge (Sadie Technology) — AI voice concierge handling reservations, FAQs, and call routing.
• Retell AI — General-purpose voice AI platform; hospitality customers for reservations and support.
• Dialzara — Affordable AI voice receptionist for smaller properties.
• Voiceflow — Build-your-own conversational AI with hospitality-specific templates.

Revenue Management (RMS / pricing)

The most mature AI category in our industry — but the fastest changing.

• IDeaS Revenue Solutions — #2 RMS by hotelier rankings; AI-powered automated pricing. HotelTechReport The enterprise standard.
• Duetto — Trusted by 7,200+ properties across 100+ countries; Open Pricing model for dynamic rate optimization across every segment, channel, and stay date. COAX Software
• LodgIQ — AI-powered commercial strategy platform trusted by 550+ hotel brands LodgIQ; recently launched AI Wizard, the industry's first generative AI platform purpose-built for hotel revenue management with conversational interface. LodgIQ
• Atomize (now part of Mews) — Fully automated AI-driven pricing in real time; acquired by Mews in November 2024. HotelMinder
• RoomPriceGenie — The #1 RMS choice for small and mid-sized properties. RoomPriceGenie
• Cloudbeds Revenue Intelligence — Causal AI engine natively integrated with the Cloudbeds PMS.
• BEONx — Total profitability platform; shifts focus from RevPAR to RevPAG.
• Beyond — Dynamic pricing for short-term rentals and hotel hybrids.
• FLYR for Hospitality — Machine learning + business intelligence (originally airline pricing tech).
• PriceLabs — Hotel/STR hybrid pricing managing 400,000+ properties globally.
• N2Pricing — Cloud-based RMS for mid-market and independents.
• Lybra Assistant — Boutique-focused RMS for European independents.
• Pace (now Open Pricer Hotels) — RMS focused on smaller hotels and groups.
• Lighthouse (formerly OTA Insight) — Market intelligence + rate shopping + business intelligence.

Have a good week.

Cheers, Sloan