AI for Mallorca businesses, built and installed.
We're a Mallorca-based team that builds and runs the AI systems small businesses in Palma and across the Balearics actually need. No complexity, no year-long project, no engineers required on your side of the table. Two or three high-value systems, built properly, then handed over or run for you.
7+ years in sales across ten industries, plus hands-on AI implementation · £1.5M in revenue tracked through the system we built for a UK-based health transformation business · Based in Palma de Mallorca · European servers, GDPR-first, EU AI Act aware
What AI can actually do for a Mallorca business
Most business owners we talk to already know AI could help them. They just don't know where to start, which tools do what, or whether the newest thing making the rounds actually applies to a business their size. So instead of talking about AI in the abstract, here are three situations any small business owner will recognize. Each is a solvable problem, and most are costing real money right now.
The enquiry that came in at 9pm
A small business owner gets an enquiry through their website at 21:47 on a Tuesday. They see it Wednesday morning. By then the enquiry has emailed two competitors and booked with whoever replied first. This isn't a lead-generation problem. It's a response-speed problem, and it's the single biggest source of lost revenue we see in small businesses.
The follow-up that never happened
A small business gets more enquiries a month than the owner can personally chase. They reply to the first message. The enquiry says "let me think about it." Nothing else ever happens. Most of those enquiries would have converted with three or four well-timed follow-ups, but the business is busy delivering, and nobody has time to chase a maybe.
The four hours a week you'll never get back
A small business owner spends four hours a week on admin. Chasing signatures, updating spreadsheets, sending invoices, moving data between tools. Nothing about that work grows the business, but it doesn't do itself either.
For a UK-based health transformation business we work with, systems that address exactly these three problems produced £1.5M in tracked revenue, 7x more sales activity from the same team, and 11,000+ automated follow-ups since launch. The section below covers what those systems actually look like.
What it can't do (yet)
Business owners we talk to usually expect one of two things. AI as a magic wand that fixes everything, or AI as an overhyped fad that fixes nothing. The reality sits in the middle, and it's worth being direct about the honest limits before you hire anyone, us included.
It can't fix a broken process. It only makes it happen faster.
If your sales process is unclear before AI, it will be unclear after AI, just automated. This is why every project starts with a short call: we look at how your business actually runs, and if what needs fixing is the process itself, we say so, even when it means we don't get hired.
It can't replace human judgement on the calls that matter.
Whether to fire someone. Whether to take on a demanding client. Pricing a bespoke deal. These are the decisions that shape a business, and AI has no business making them. What it does well: freeing up the hours you're currently losing to admin, so you have more time and headspace for the calls only you can make.
It can't work without limits, and the limits are the whole point.
An AI agent works reliably when its job is narrow ("respond to enquiries with these types of answers, book calls into these time slots"). It fails when it's given open-ended tasks ("figure out how to grow this business"). Every system we build has hard boundaries and clear escalation paths for anything outside them. The goal is quiet reliability, not autonomous magic.
It's not always right, which is why nothing important runs without a human check.
LLMs occasionally hallucinate. Automations occasionally break when an upstream tool changes. Anything that matters (a message going to a real customer, a decision affecting revenue) has a human in the loop, always. You see every message that goes out, and you can change or switch off any part of the system at any time.
If you were hoping to hear "AI can do everything and it'll pay for itself in a week," this isn't the page for you. The reality is quieter, and more useful.
Where most Mallorca businesses start
Every business we work with begins with two or three specific systems, not a strategy document. The three below cover most of what a small business owner is losing money to right now. Pick one, and it usually pays for itself before we build the second.
Speed to lead
The system: an enquiry hits your website, WhatsApp, or contact form. Within seconds (at 3am, on a Sunday, when you're on holiday) the enquiry gets an intelligent first response tailored to what they asked about. It answers their question, qualifies whether they're a fit, and books a call straight into your calendar.
Best fit: any business where enquiries come in outside working hours, or where the founder is the one replying and the founder is busy. Which is most businesses.
Follow-up automation
The system: the enquiry that said "let me think about it" enters a sequence, across email, WhatsApp, or SMS, whichever channel fits. It keeps in touch personally, referencing what they originally asked about, until they either book or explicitly opt out. Not a mass email. An actual sequence tailored to their conversation.
Best fit: for businesses that get more enquiries than they can personally follow up on, where "I'll circle back to that" is a familiar internal apology.
AI receptionist
The system: an assistant, embedded on your site or handling your inbox, that fields the common questions ("Do you take new clients?" "What are your prices?" "Where are you based?") and passes anything that needs judgement to a person. It's on 24/7, doesn't get tired, and never forgets to reply.
Best fit: for businesses fielding the same handful of questions dozens of times a week, where a person's time is better spent on actual work.
Most businesses start with one of these, prove it works, then decide whether the next one is worth doing. Nothing is "phase one of ten." One system at a time.
European rules, in plain English
Data privacy and AI regulation are the two questions that come up on every first call with a European business owner considering AI implementation. Both have straight answers. Here's what actually matters.
GDPR: where your customer data lives, matters
Under GDPR, you're responsible for what happens to your customers' personal data, including where it's stored and who processes it. When you use a US-based AI service, that data typically leaves the EU, which puts the compliance burden on you to document and defend.
Our approach: everything we build for you runs on European servers (our own, in France). Customer data stays in the EU. Nothing gets shipped across the Atlantic to make an AI tool work.
The EU AI Act: most small-business use is low-risk
The EU AI Act is the world's first serious AI law. It categorizes AI systems by risk level, and the good news is that what you'll actually use (chatbots, follow-up sequences, receptionist agents, workflow automations) sits firmly in the "limited risk" or "minimal risk" tiers, where the requirements are light. Being aware of them is enough for the vast majority of small-business use cases.
Where it starts to matter: automated hiring decisions, credit scoring, biometric ID, exam scoring. These are high-risk categories most small businesses don't touch.
What that means practically
Our principle: anything customer-facing includes clear disclosure that AI is involved. Clarity is a starting point, not a compliance chore. We don't build systems in the "high-risk" or "prohibited" EU AI Act categories, and we'd tell you if what you asked for fell into either. Every system logs what it does, so you can produce an audit trail if you're ever asked.
This isn't legal advice. For that, talk to a lawyer. But it's the practical starting point most small-business owners are looking for, and it's the shape of every project we run.
How it actually works
Most AI implementation projects start with a proposal and a six-month contract. Ours starts with a 30-minute call and one system built. Here's what actually happens if you get on that call, from the moment you book to the moment something is live and running.
Step 1: A 30-minute call, no commitment either way
We look at how your business actually runs. Where enquiries come from. How you reply. Where things sit longer than they should. What's already automated, and what's still being done by hand at 11pm on a Sunday.
At the end of that call, you'll have a specific list of the two or three systems worth handing over first: the ones that would return the most time or the most missed revenue, in the shortest time. That list is yours to keep whether you hire us or not. Some people take it and DIY. Most decide the DIY route isn't worth the six months, and hire us.
Step 2: We build one system, properly
If you decide to work with us, you pick which of the two or three to start with, usually the highest-pain one. We build that one system end-to-end, connected to the tools you already use. Timelines vary by system, but "weeks, not months" is the honest range for a first install. Not because we cut corners, but because a well-scoped first system doesn't need a year.
Nothing else in your business changes while we do it. You keep running the way you always did; the new system comes alongside, not in place of.
Step 3: You see it work, then decide what's next
The first system runs. You see it in your numbers: bookings that came through overnight, follow-ups that closed a maybe, hours you got back. From here, you can run the system yourself, or we can run it for you as an ongoing service. Either way, once it's proven what it's worth, you decide whether to add the next one. Or don't. You go as far as you want, and no further.
No lock-in contract. Each next system is a decision you make with results in front of you, not projections in a slide deck.
Proof, not promises
Two proof points, both of them ours. One is a client, the other is our own business. All numbers below are measured in real systems, not estimated.
Proof 1: A UK-based health transformation business
Before: everything ran on spreadsheets and memory. Enquiries got a reply only when someone remembered. Most were never followed up. There was no way to see which deals were being won or lost.
What we built: a full CRM (every enquiry moved off spreadsheets), instant automated replies to new enquiries, follow-up sequences that kept going until someone answered, and live reporting on what was closing and what was slipping.
The measurable outcome, pulled directly from their own CRM:
£1.5M in revenue tracked since the system went live
390 deals closed and tracked, where previously none were even visible
11,000+ follow-ups sent automatically since launch
7x more sales activity from the same team, with no new hires
Every figure above is measured, not estimated. Happy to walk through the full case in detail on a discovery call.
Proof 2: The business you're reading this on
Viacala runs on the same stack we install for clients. The chat widget on this site is powered by an AI agent that fields common questions, qualifies visitors, and passes real opportunities to a person. The 6-question audit is a full lead-capture pipeline built in n8n. The dashboards we use to run our own business are built on Supabase, with the reporting layer in Metabase.
Every one of those systems is something we'd build for a client. We tell you exactly how we built each one in our blog posts.
If you want to see the systems working before you hire us, they're working right now on this site.
That's the whole proof: one client's numbers, and a live demonstration you're currently interacting with. Everything on this page you can verify by clicking around the rest of the site. If neither is enough, no proof will be.
What it costs
Most agencies quoting AI automation work start around €10k+ for a first project. Enterprise AI consulting runs several times that. Ours usually costs less than either, not because we cut corners, but because we scope tight, don't bill by the hour, and don't add layers of project management to work that doesn't need them.
The rough shape
We quote a fixed price for the build itself, based on what the audit uncovers. Most first projects fall meaningfully below that €10k agency floor. If you want us to run the system for you as an ongoing service (rather than handing it over), that's a monthly rate we agree upfront. If you'd rather run it yourself once it's built, there's no monthly cost. You own it.
Every quote is fixed. You'll know exactly what you're paying before we start, and it doesn't change unless the scope does. No hourly billing, no "creative overages."
What moves cost up or down
More expensive: the system touches more tools · custom logic that doesn't exist in off-the-shelf tools · higher volumes (agents handling thousands of interactions per day, not dozens) · larger historical data to migrate in.
Less expensive: a clear process that just needs automating (not redesigning) · modern tools you already use, the kind that integrate cleanly with automation platforms · starting narrow and expanding, rather than trying to boil the ocean.
In most cases the audit call gives you a clear fixed price for the first system on the spot. You'll get it in writing after, but you'll never wait days or weeks for a proposal to arrive.
Payment terms
Standard, discussed and agreed on the discovery call. Nothing unusual, nothing surprising.
The point isn't to be cheapest. It's to be honest and predictable, so you can budget for it and know exactly what you're getting for what you're paying.
Questions before you book
How quickly can we get started?
The discovery call usually happens within a week of you booking. If you decide to work with us, the first system build starts within another week or two. Most first systems are live within four to eight weeks of the discovery call.
What happens if the first system underperforms?
The first system is scoped small on purpose, so underperforming isn't a major setback. If something's broken, we fix it at no extra cost. If the assumptions were wrong, we adjust the scope. What doesn't happen: you paying for a system that never works.
Can we start with a smaller pilot?
Yes, in some cases. If a first project has meaningful uncertainty, we can scope a lower-cost pilot to prove the approach before committing to the full build. Not every project needs one. Most first systems are already small enough that a pilot would just add steps.
Do you work with businesses that operate in Spanish or German?
Yes. Our client-facing communications are in English, but the systems we build can operate in Spanish, German, or any language your customers actually use. The AI models we work with are natively multilingual.
What if AI tools change and my system becomes obsolete?
The tools we build with are in active development with backward compatibility as a priority. Substantial changes sometimes need updates, as with any software. Our ongoing service option covers this; if you're running the system yourself, we're happy to advise on updates when they're needed.
Do you work with businesses outside Mallorca?
Yes. We're based in Mallorca but work with clients across the UK, Spain, and the wider EU. Delivery is remote (video calls, cloud-based systems), so location doesn't limit us. Being in the same time zone is enough.
Will AI replace my team?
No, and it shouldn't. We automate the repetitive work no one enjoys or should be doing by hand. Your team keeps the work that needs judgement, relationships, and experience, and gets more hours to do it properly. Most business owners find their team ends up handling more clients, not fewer people handling the same.