We assess, design and implement agent systems and automations into your business's day-to-day operations. Assessment and strategy are part of the work; what you get at the end is the system running and your team using it.
We work in Spain and across Europe with all kinds of businesses that want to grow, whether you're a one-person business or you have hundreds of employees. If your process won't deliver a clear return, we tell you before we invoice you for anything.
Take your free AI assessment →100% freeNo commitmentTakes 2 minutes
An artificial intelligence consultancy analyses how a business works, identifies which processes can be carried out with AI, and designs and implements the systems that do it. The complete job has five parts:
Assessment. Understanding how you really work today and what each process costs, in hours and in money.
Solution design. Deciding what gets automated, what gets handed to agents, what gets integrated and what isn't worth touching yet.
Build and integration. Building it and connecting it to the tools the business already uses.
Team training. Making sure the people who'll be working with the system know how to use it from day one.
Monthly service. Keeping it running and up to date when an API, a process or the underlying AI model changes.
Most consultancies deliver the first two: the assessment and the design. That's legitimate and can be worth a lot of money, but on the last day you go home with a document. We deliver all five, and that's why on the last day there's something up and running.
So what gets automated in practice? Almost always the same areas: invoicing and accounting (reading invoices, extracting data, getting it ready for the ERP), moving information between systems (from email to Excel, from Excel to the CRM), first-line customer service (sorting what comes in, answering the repetitive questions, escalating the rest), recurring reports, onboarding and offboarding customers, employees or suppliers, sales follow-up and document management. None of these processes gets automated just because it's on the list: it gets automated if, in your business, it happens a lot, follows rules that can be written down and works with information that's already in some system.
If you want to know more, we go into it in detail in this article: which business processes to automate.
Reinicia does all five because the last three are where most AI projects fall apart: a plan gets approved, nobody integrates it with the real tools, the team doesn't use it and three months later the business concludes that "AI wasn't for us".
There's no way to tell from the outside whether an AI consultancy is any good. There is a way to tell whether the project is well thought out: these three questions reveal it in five minutes, and you should get the answers in writing before you sign.
What exactly do you hand over on the last day?
If the answer is a document, a roadmap or a training session, you are buying advice. It may be worth it, but don't expect anything to change on its own.
Our answer: the system running inside your operations, documentation on how it works and your team knowing how to use it.
Who maintains it when something changes?
Every system connected to real tools breaks at some point: an API changes, a form changes, an internal process changes. And there are two things almost nobody tells you before you sign: AI costs money (tokens, APIs, subscriptions) and models become obsolete in about six months. A system nobody updates degrades on its own.
Our answer: a monthly service. It covers the servers your system runs on, ongoing updates to the AI models, keeping your agent systems and automations up to date, and the AI costs spread across the fee so you don't have to deal with APIs or subscriptions. Changes and new features are quoted separately, and we tell you that upfront, not after the fact.
How will we measure whether it's worked?
If there's no figure before you start, there won't be one afterwards either. We measure the process as it is today (hours, volume, errors, cost) and measure it again once the system is running. If the improvement doesn't show up in the numbers, it never happened.
Not every business needs the same thing, and they don't always need the most expensive option. This is what we do, when it makes sense and when it doesn't.
Whether you know exactly what you want or you just know something doesn't add up, tell us about your case. That conversation tells us which of the four is right for you (assessment, automations, agent system or training), in what order, and also whether none of them is right for you yet.
Tell us about your case →Four stages, always in the same order. Between the first and second stages there's a way out: if the numbers don't work (if what you'll save doesn't pay for what it costs), we tell you and there's no project.
We analyse how your business works: processes, tools, bottlenecks and where the time goes. We come out of this stage with a figure for what the process costs today, not a hunch.
We turn the assessment into a concrete solution: what to automate, what to hand to agents, what to integrate and what isn't worth touching yet. This is where the decision is made, with a fixed quote in writing.
We build the system and implement it inside your operations: agents, automations, connections to your tools, documentation and training so the team uses it from day one.
You measure time saved, errors reduced and capacity gained. The system is documented, the team knows how to use it and we make sure it keeps working and stays up to date for as long as it runs.
Spanish Tax Agency requests: from 24 h to 2 h per process.
JGM Consultores case
Off operating costs: from €480 to €10 per process.
JGM Consultores case
At Grupo Platea, for work that used to take 4-5 days of review.
Grupo Platea case
A Spanish Tax Agency request used to take them about three working days, around 24 hours. Today it's resolved in about 2 hours. The operating cost of the process went from about €480 in staff hours to about €10 in AI costs.
Human review of their automations used to take between 4 and 5 working days. Today it's around 1 hour a day, with far more control over what goes out.
We've implemented a complete agent system that searches for European project tenders and drafts the proposals. Submitting a tender currently takes them between 3 and 4 weeks of work, which caps them at around 16 tenders a year. With the agent system they expect to be able to bid for between 50 and 60 tenders a year, and they expect the company's turnover to triple from this system alone. These are forecasts, not final results: we're measuring them with the client and the real figures will be published here when we have them.
These are figures from specific clients, not a market average or a projection. Your process could deliver more or less: that's why the first thing we do is measure it as it is today, before touching anything.
We don't publish rates, because the price of an implementation depends on the process, the volume and the tools you already have, and a ballpark figure on a website only ends up misleading someone. What we can publish is how you pay, because that's what really breeds mistrust when nobody spells it out.
Half before the implementation starts and the other half when the system has been delivered and is working. No exceptions, and we say so in the first meeting.
The assessment produces a specific scope with a fixed price. If something comes up during the project that changes the scope, we talk about it beforehand, not on the invoice.
The systems run on our servers and are kept going by a monthly service. There are two things worth knowing before you sign anything: AI costs money (tokens, APIs, subscriptions) and models become obsolete in about six months. The monthly service covers the servers, ongoing model updates, keeping your agent systems and automations up to date, and the AI costs spread across the fee so you don't have to deal with APIs or subscriptions yourself. Changes and new features are quoted separately.
We don't do cheap packages or hours here and there. If your priority is the lowest quote on the market, there are cheaper options than us. We won't go looking for them for you, but tell us from the start and neither of us will waste our time.
Behind Reinicia is its founder, Olmo Romero. He has spent more than 20 years automating processes (he started well before ChatGPT existed) and 16 years as a trainer for European institutions, businesses and master's programmes, among others. He set up two companies before this one and Reinicia is the third: when he tells you whether a process is worth automating or not, he says it as a business owner, not as a supplier.
The person who assesses your case is the same person who designs the solution and who answers when something goes wrong. When more capacity is needed, it comes from the agent system Reinicia uses to run its own business: the same one that is then installed for clients. We sell what we use every day.
It analyses how a business works, identifies which processes can be carried out with artificial intelligence, and designs and implements the systems that do it. The complete job has five parts: assessment, design, build and integration, team training and a monthly service to keep it running and up to date. Most consultancies deliver the first two; we deliver all five.
It depends on the process, the volume and the tools you already use, so we don't publish rates: a figure quoted without knowing your case wouldn't be honest. What is fixed is the model: a fixed quote before we start, 50% at the start and 50% on delivery, and then a monthly service while the system is running. That monthly service covers the servers, ongoing updates to the AI models and the AI costs spread across the fee, so you don't have to deal with APIs or subscriptions yourself.
It depends on the scope. Some implementations are quick, a few days or a week, when the process is clear and well defined. Others take several weeks, when several systems need connecting or the work calls for judgement and not just steps. What is never instant is the groundwork: understanding how you really work. In the assessment we give you the timescale for your case before you decide anything.
No. We work with all kinds of businesses that want to grow: whether you're a one-person business or you have hundreds of employees. Size matters less than two things: that there is a specific process with real volume, and that there is someone with the authority to decide.
No, and the difference is huge. Knowing how to use ChatGPT means that you know how to ask AI and it knows how to answer you. Nothing more: someone on your team has to be sitting in front of it, asking, every single time. What we implement are agent systems capable of carrying out complex tasks from start to finish inside your processes: they read, classify, draft, look things up, decide within the rules you set and coordinate with one another. They can take over the work of one person or several. Having ChatGPT accounts isn't adopting AI: it's buying a tool.
An automation runs a defined workflow: always the same steps, in the same order, with no room for manoeuvre. It's reliable and cheap when the rules are stable. An agent system is a different kind of thing altogether: it is the equivalent of a worker. A virtual worker with all its capabilities (it reads, interprets, drafts, looks things up, decides the next step within the rules you set and knows when to ask you) or, if the work calls for it, an entire team of agents, each an expert in a different area, coordinating with one another to carry out complex tasks. With an automation you buy a workflow; with an agent system you buy working capacity. We use each approach where it adds value: if a simple automation solves your case, we won't sell you agents.
That's decided with you before anything is built, and it's part of designing the solution. We define which information leaves your business environment and which doesn't, what gets anonymised, where each step is logged and who reviews before anything is signed off. In sensitive processes (tax, accounting, customer data) human review isn't removed: it's freed from the mechanical work.
We're based in Madrid and work remotely for almost the whole project, with in-person meetings when they add something: usually the kick-off and the team training. We have clients in Spain and Belgium, and we've delivered training in Brussels and Venezuela. Distance has never been a problem.
Yes, and it's included in every implementation: a system nobody knows how to use is no use to anyone. We also offer training as a stand-alone service, for teams or for management, built around your own processes rather than generic examples.
We tell you and there's no project. It has happened before and it will happen again. We'd rather lose a project than charge for an implementation we know won't pay back what it costs: Reinicia's business is the clients who are still with us in year two, not the ones who sign once.
Describe the case and we'll tell you frankly whether the numbers work or not. If we see a clear case, the next conversation will already have a specific proposal on the table.
Let's talk →No commitment. We only suggest a call if we see a real fit. We reply within 24-48 working hours.