There is a line repeated so often in technology pitches that it has lost its edge: "artificial intelligence isn't here to replace people, it's here to empower them." We have used it ourselves. It sounds good and is generally true. The problem is that as an argument it commits to nothing. Anyone can say it and no one can verify it. When it comes time to hire, it doesn't tell you what you will actually get.
It is worth pausing on what it really means. Empowering a person isn't placing a language model beside them and hoping they perform better. It is giving back three things the job takes from them every day: time, certainty and peace of mind. Time, because a good part of a team's day goes to repetitive tasks, to reports that take days to assemble and to manual processes no one questions because they have always been done that way. Certainty, because when the information finally arrives, often no one manages to verify where it came from before deciding with it. Peace of mind, because almost everything a company does today carries an associated rule to comply with. Working without knowing for sure whether you are complying is a quiet form of wear and tear.
That is where we saw the opportunity, and the answer was not to build an AI. It was to look at three concrete problems and solve each one with the tool it calls for.
Three problems, three solutions
The first is about analysis. In the financial system and in large companies, the people who decide do not access the data directly. The board asks for it, the request goes down from the manager to the lead and from the lead to the analyst, then comes back up reviewed. Weeks can pass. And when one more pass is needed, you wait again for the analyst to have time to build it. Alicanto Analysts tackles exactly that. It is a conversational agent that connects to the organization's internal databases so you can ask it in plain language and get back prose, charts and the detailed process that was run to obtain the result. Anyone can try it free on public Central Bank and CMF data before connecting their own. It is designed to abstain when it doesn't have the data rather than invent it, because a figure with no verifiable origin isn't information but an opinion that looks like a fact.
The second is about compliance. Law 21.719, on personal data protection, requires companies to handle that data to a standard that didn't exist before. Companies will need consents they can prove, data subjects will have rights that must be honored within the legal deadlines, and security breaches will have to be reported to the regulator. For that we built a compliance platform whose core is deterministic. Before the law, what matters is not that a system looks intelligent, but that every step is proven with absolute certainty. That is why the core of this solution is automation and traceability, not the output of an agent that might improvise. Each consent is versioned with the exact text the person accepted, each data-subject request runs with its 30-day legal deadline, and each step is sealed into a chain of evidence that can later be audited.
The third is about management. Ley Karin, Law 21.643, requires every company to provide a channel for reporting workplace harassment, sexual harassment and violence at work. Each report has deadlines in business days, protective measures for whoever reports, and mandatory referral to the Labor Directorate. Its core obeys a similar demand: what matters is meeting the deadlines, keeping the reporter's confidentiality and being able to prove that each step of the case file was done when it should have been. Our whistleblowing channel tracks those deadlines with a due-date traffic light and keeps the case file in a record that cannot be altered after the fact.
The right tool, not the fashionable one
In none of the three solutions is artificial intelligence the protagonist, and that is deliberate. We see it as one more tool, not as our identity. We integrate it where it genuinely adds value, not where it sounds good to say we use it. The temptation of the moment is to slap "AI" on everything because it sells, and that is precisely the shortcut that ends up promising magic and delivering risk. We prefer the opposite: choosing the tool by the problem and not by the label. Sometimes that is a conversational agent. Sometimes it is automation that runs the same way every time. And when no standard solution fits, we do custom work that combines on-demand process automation with consulting that interprets the regulatory and technical demands.
That the core of Protect and Manage is deterministic doesn't mean artificial intelligence is left out. It comes in where it adds value without putting traceability at risk. Today we use it heavily in onboarding, so that getting a new company's compliance up and running takes days and not months. And we are building the modules that bring those platforms the same architecture as Analysts: so that every organization can ask its own compliance data in plain language, get actionable signals and always see the source behind each answer. Those modules are still in development; AI-assisted onboarding is already part of how we work. The tool changes with the case. The purpose is always the same: to give you back time, certainty and peace of mind.
That is also what makes coherent three solutions that, seen from the outside, look like they belong to different worlds. In all three, the center is neither the model nor the automation, but the evidence. That you can open any number and see where it came from. That you can prove to a regulator or to your own board that every step was done as it should have been. In Analysts that evidence is the source of the data you will decide with; in Protect and Manage it is the proof that the company is complying with the rule. That proof is what lets you get on with your business instead of living on edge about enforcement. A technology you can't open and show in full doesn't help you decide: it only asks you to believe it. And believing blindly is exactly the risk we are here to take off your shoulders.
This is the beginning. We will get some things wrong and learn along the way, and we will tell it here without dressing it up. But the starting point is already set and it fits in three words you will see us repeat: analyze, protect, manage.
