# Sandra Kowalczyk, Owner/Operator at The Heron House — read of HotelAI, May 11 2026

> 11 years running a 48-room inn on the Door County peninsula. Came from brand management, which means I think in funnels but swear in hospitality.

## How I got here

Tuesday morning, between the 9 AM checkout rush and a linen delivery. I was scrolling the Wisconsin Lodging Association Facebook group (yes, still, I know) and someone dropped this link in a thread about "tools that actually replaced a hire." Four people had liked the comment. That was enough for me to tab over. I've looked at Whistle and Canary before and got as far as a demo call with each. Neither stuck.

## What I clicked first

The pain section, not the hero. I skimmed past "Your front desk never sleeps" because I've read that line on four different tools. What actually stopped me was this: "your team spends more time on coordination than connection." That's the exact sentence I said to my husband last October when we were talking about whether to add a third front desk person. I didn't click anything, I just read the rest of the page.

## Where I paused

The review management module. Specifically: "HotelAI drafts personalized responses to new reviews within the hour they post, across Google, TripAdvisor, Booking.com, and Expedia." I do this myself on Sunday nights. It takes 45 minutes and I resent every second of it. I've tried to delegate it once and the response my desk person wrote started with "We are deeply sorry for any inconvenience." I published it anyway because it was 10pm. The fact that this product calls out "approve once and publish across all platforms" and mentions a threshold rating for auto-publish tells me someone who actually runs a property wrote this feature spec.

## What I distrusted

Three things, in order of how much they bugged me.

One: "Marcus L., Owner, 34-room boutique inn, Vermont." No property name. No link. No photo. The quote itself is good -- "the answers it writes are genuinely better than what we were writing by hand" -- but I can write that quote. Anyone can write that quote. Where is Marcus? What's the inn called?

Two: "92% Pre-Arrival Response Rate." I had to re-read this twice. Response rate from whom? If HotelAI is sending automated messages, of course they're delivered. That's not a response rate, that's a delivery rate. I wanted to see guest reply rates, or conversion to upsell, or something that means what it sounds like it means.

Three: There's something on the page that says "Unlock the dossier -- $5" and also navigation items that read "For FAQ Pricing Honest Hire team to build." I genuinely could not figure out what that was. Is this a product or a portfolio piece? That confusion made me slightly less confident in everything else I was reading.

## What would convince me

One real property name attached to Priya's Austin result. Not "78-room independent hotel, Austin TX" -- the actual hotel name, so I can go look at their TripAdvisor profile and see if their score actually went from 4.1 to 4.8 in three months. I can verify that in four minutes. If that's real, I'm booking a demo today.

Alternatively: a screen recording of the actual dashboard in use at a real property. Not a slick product video with mockup data. Something that looks like someone's actual hotel with their actual guest threads and real review drafts queued up. The "Watch the 30-second explainer" link was on the page but I'm working from text here so I don't know if that video delivers.

## What I'd ask in an email reply

1. The Priya case study -- what property is that, can I call her or look them up? A 0.7 score gain in three months is either the most impressive thing I've seen or a number that needs a lot of footnotes.

2. How does the "set your voice and policies" step actually work for a property like mine? We're not a boutique brand with a 12-page guide document. Our "voice" is whatever I sound like when I'm being nice after a long shift. Has anyone onboarded a 2-person front desk operation where the brand guide is "be warm and don't overpromise"?

3. What happens when the AI gets it wrong? Specifically: a guest messages at 2am that they locked themselves out and there's a dog in the car. Who does HotelAI contact, how fast, and what does that escalation actually look like in practice? That's the scenario I care about, not the 92% pre-arrival rate.

## Verdict: curious-enough-to-reply

The review management feature and the specific framing of the problem are better than most tools I've seen in this space. But I need one real, verifiable proof point before I spend any more time on it. If someone replies to my questions with Priya's hotel name and a 3-minute Loom, I'm going to a demo call.

---
*Memo by skeptic persona, generated 2026-05-11. Studio breaks own self-grading loop.*