HiJiffy review: WhatsApp automation that actually works

I run three boutique hotels across England, about 120 rooms between them. We’re a small collection, not a chain in any corporate sense. Guests expect to feel looked after without being fussed over, and my teams do everything from check-in to stacking the breakfast dishwasher. I trialled HiJiffy at the Cotswolds property first (45 rooms, the smallest of the three). Any technology I bring in has to prove itself at one hotel before I consider rolling it out to the other two. Every hour spent learning a new dashboard is an hour not spent making sure room 14’s radiator isn’t making that noise again.

HiJiffy came onto my radar because three hoteliers I trust in Portugal and Spain all said the same thing: “the WhatsApp piece is good.” I’m suspicious of consensus, so I signed up for a trial.

WhatsApp is the bit that matters

Let me skip to the thing that justifies the subscription. HiJiffy connects to WhatsApp Business API and lets guests message your hotel the same way they’d message a friend. They ask about parking, check-in times, whether the restaurant is open on Mondays. The AI responds instantly, accurately, in whatever language the guest wrote in.

I was sceptical. My guests are mostly British, with a good number of Americans and a growing share of French and German visitors. The AI handled all four languages without any trouble I could detect. A German guest asked about walking routes near the hotel (in German) and got a sensible reply within seconds. My receptionist would have taken five minutes to look that up and compose a reply. Multiply that by thirty similar queries a week and you start to see the point.

The pre-arrival messaging is where the open rates get interesting. We used to send a pre-arrival email. Maybe 40% of guests opened it. HiJiffy sends a WhatsApp message instead. Our open rate jumped to somewhere around 85%. Guests reply to it, too. They ask about early check-in, restaurant bookings, directions. The AI handles most of these without any staff involvement, and the ones it can’t answer get routed to my front desk with the full conversation attached. My team sees what’s already been said and picks up from there. No guest has to repeat themselves.

I was curious about what larger properties were seeing, so I did some asking around. A colleague who consults for Hotel Sacher in Vienna mentioned their chatbot automates 96% of guest queries. Leonardo Hotels, which is a rather different operation to my three small hotels, reported 93% automation across 281,000 queries. My numbers are lower (I’d estimate 70-ish percent at the Cotswolds property), but I also have fewer repetitive questions to begin with. The point is that the AI holds up at scale, which matters if you’re thinking about this for a bigger property or rolling it out across several.

The chatbot beyond WhatsApp

HiJiffy also puts a chat widget on your website. It handles booking queries: a visitor asks about availability for certain dates, the chatbot checks, presents options, and nudges toward a direct booking. We’ve tracked a handful of direct bookings per month that came through this route. Not life-changing revenue for a hotel our size, but it pays for the subscription several times over.

The chatbot’s knowledge comes from a setup process where you feed it your hotel’s information. Room types, amenities, policies, local recommendations. This took me about two hours, which felt like a lot, but once it’s done the AI uses everything contextually. A guest asks “is there parking?” and it knows we have a free car park behind the building for up to 20 cars. That specificity is what makes it feel less like a bot and more like a well-briefed concierge.

I should be honest about the learning curve, though. Their support team is good, and they clearly know the onboarding process inside out. But the bot itself takes longer to train than you’d expect. After the initial two-hour setup, I spent another couple of weeks correcting answers and adding edge cases before it stopped giving vaguely wrong responses to questions about our afternoon tea service. One hotelier I spoke to put it more bluntly: “long process before the bot understands our guests.” That matches my experience. Budget a month before it feels properly dialled in.

There’s also a thing with canned responses. Any template or pre-written reply needs pre-approval before it goes live, which makes sense from a quality control perspective but can slow you down when a situation is urgent and you just want to send a quick message. I hit this twice during a busy weekend and ended up replying from my personal phone instead, which somewhat defeats the purpose.

Where the AI stumbles is with unusual or compound requests. “Can you arrange a birthday cake for my wife and also move our dinner reservation to 8pm and by the way is your pool heated?” That kind of message gets a slightly confused response, or the bot answers the first question and ignores the rest. My staff catch these because HiJiffy flags when confidence is low, but keep in mind the AI isn’t infallible with messy, multi-part messages.

But it’s not just compound messages. Even straightforward questions get answered wrong sometimes. A guest asked what time our restaurant opens on Sundays. The information was in the system, correctly entered. The bot gave the weekday hours. That’s not an edge case. That’s a basic lookup that should work every time, and it didn’t. The more I used HiJiffy, the more it felt like well-packaged automation rather than properly implemented AI. It follows rules and matches patterns, but when anything falls slightly outside the expected flow, it guesses rather than asking for help.

The WhatsApp channel itself is also more limited than I expected. Rich media support is thin. Buttons, interactive lists, product catalogues, the kind of components you’d use to build a proper pre-arrival experience, are either missing or restricted. When I tried to build a more detailed pre-arrival message with check-in options, local recommendations, and an upsell offer in one flow, I hit the limits quickly. You end up sending multiple plain text messages instead of one well-structured interactive message. For a tool that’s built around WhatsApp, the range of what you can actually do inside a WhatsApp conversation feels narrow.

The setup headache

Getting WhatsApp Business API working requires Meta verification. I’ll be blunt: this was a bureaucratic ordeal. You need a verified Facebook Business account, you need to submit documentation, and Meta’s approval process moves at a pace that suggests nobody is in any particular hurry. HiJiffy’s support team helped, and they clearly know the process well, but it still took about two weeks before we could send our first WhatsApp message. If you’re expecting to sign up on Monday and go live on Tuesday, adjust your expectations.

The dashboard itself is ugly, and I don’t say that lightly. I’m not someone who cares much about how software looks, but HiJiffy’s interface makes simple tasks harder than they should be. The navigation steps aren’t logical. Finding the analytics section takes three clicks when it should take one. Setting up a new automation involves jumping between screens in an order that makes no sense until someone explains it to you. After a few weeks you learn where everything lives, but the first week involved a fair amount of muttering and the second week wasn’t much better.

Features I don’t need

HiJiffy sells to hotels much larger than mine, and it shows. There are campaign tools, segmentation options, and automation workflows that a 200-room city hotel would put to good use. For a 45-room property like the Cotswolds, half of these features sit untouched. Some of the campaign tools make more sense once you’re thinking across multiple hotels, but even with three properties I’m not using the segmentation features. I’d prefer a version of the pitch calibrated for smaller operators. Tell me about WhatsApp and the chatbot. Skip the enterprise automation deck.

European credentials

HiJiffy is headquartered in Lisbon. They’re a European company, funded in Europe, building for European hotels. The data stays within EU jurisdiction. Marc would approve, and Anna would find their GDPR setup reasonable (though she’d probably still find something to question). For my purposes, it’s reassuring to know that guest data isn’t sitting on a server in Virginia subject to American law.

One thing worth knowing: HiJiffy’s AI runs through what they call Aplysia AI, which uses multiple large language model providers underneath, including OpenAI, Anthropic, Google, and Mixtral (the open-source model from French company Mistral). That multi-provider approach is unusual in this category. Most competitors just plug into OpenAI and call it a day. The fact that HiJiffy includes a European AI option (Mixtral) alongside the American ones at least shows they’re thinking about the question, even if guest data still touches US-based AI processors for some queries.

Since my first look, I’ve done a bit more homework on the company itself. They now serve over 2,500 hotels across 60+ countries, which is a reasonable footprint for a company that’s raised about €6M, mostly from Portuguese and EU investors. If you’re a small hotel owner wondering whether they’ll still be around in three years, that trajectory is encouraging. Not a guarantee, but encouraging.

Their support team is responsive. I’ve had questions answered the same day, and when I had a technical problem with the WhatsApp connection dropping, they fixed it within a couple of hours. Support in English was perfectly fine, though I imagine Portuguese and Spanish speakers get an even better experience.

Is it worth it for a small hotel?

A note on pricing: HiJiffy doesn’t publish their rates, which is one of my pet peeves with hotel tech vendors. Sophie would be furious. You have to go through a sales call to get a quote, and the final figure depends on your property size and which features you need. From what I’ve gathered from my own contract and conversations with other hoteliers, expect something in the range of €3–5 per room per month, though the starting point for smaller properties seems to be around €149/month. For the Cotswolds property (45 rooms), the maths works out to roughly €180 a month. Across all three hotels it’s a bigger line item, but the savings scale too. The direct bookings through the chatbot more than cover it. The time saved at reception is harder to quantify, but my front desk manager at the Cotswolds told me (without being asked) that she spends less time answering repetitive WhatsApp messages since we set this up. When staff voluntarily praise a piece of technology, that’s the most reliable review you’ll get.