Our approach

It is not about the ticket. It is about the context.

The problem almost everyone knows

You probably know the situation. You explain your business to a technical team. Hour after hour, every detail. And still the feeling remains that nobody has really understood what this is about. More meetings follow. Mistakes follow. Rework follows.

At some point a quiet reluctance sets in to change anything about your software and processes at all. Not because you do not want to, but because the people who are supposed to help you never quite grasp what you need. And this is not about a single problem, it is about your entire domain.

Understanding the problem is not enough

A typical ticket reads: “I need field X in form A so that I can carry out action Z.” Treating that ticket in isolation and building exactly what it asks for falls short.

The decisive question is why that field is needed in the first place. Deliver only what was asked for on the surface and you produce exactly the outcome described above: a service that was paid for and that nobody can work with in the long run.

Understanding the context is the key

We work differently. A solution only works if you understand the problem in depth and take in the whole context: terms and vocabulary, subject matter and expertise, working processes, the interests of your clients, the existing software, and your available time and people.

So we question the actual goal behind the ticket. “I need field X in form A” suddenly turns into a very different question. And that is when room opens up for the insight that you may not need the form at all to reach your goal. That there is another way, a better, faster and cheaper one.

A room full of boxes

An example, to make this tangible. Picture someone living in a single large flat. One room, a great many things, most of them stored in boxes. And that is exactly the problem: he never finds what he is looking for.

He calls in an IT specialist. The specialist enters the room, sees boxes and has no idea what is inside them. He does not even know what a box means in this context. He knows nothing about living spaces, and nothing about what these things are used for. So he asks: what do you need from me?

The answer is: more room on the labels. The current ones are too small to describe what is inside. The specialist delivers bigger labels, perhaps even a digital search on top. Ticket closed, problem solved at first glance. And the resident carries on living in one large room full of boxes.

This is where we do not stop

Our expert enters the same room and knows just as little about the boxes and their contents. He hears the same answer. But he does not settle for it, he keeps asking: what are the boxes for? Why do you need to know what is inside? What do you do with these things?

And the resident begins to explain. With a box labelled “kitchen equipment” he at least roughly knows where to look. But when he bakes biscuits he needs particular utensils at hand quickly: cutters, a rolling pin, special trays. So he set up a second, smaller box with duplicates from the large one. Over the years more equipment arrived, the large box became harder to survey, and so further specialised boxes appeared for further occasions.

That makes visible what this is really about. Not the labelling of the boxes. It is about how things have to be organised so that particular tasks can get done.

The solution that follows from it

At this point our expert says: your real problem is not the label. Your real problem is how your things are structured so that you can find them quickly for the tasks you have.

That insight leads to an entirely different solution: rooms with clear functions. A kitchen for all the kitchen equipment, a living room for social interaction, a bedroom for rest. Things sit where they are used. Suddenly no box has to be searched, there are no unnecessary duplicates, and everything is organised along actual use.

And because the domain is understood, the thinking can go further. Anyone who knows that their client lives in Germany also knows from that context that smoke detectors are a legal requirement there, and can point it out before it turns into a fine. That is how proposals arise here that nobody asked for and that still make the biggest difference.

The catch

For us to build that context at all, you would in principle have to invest an enormous amount of time. You would have to explain in detail how your world works: your line of business, your processes, your software, your entire domain.

Large companies keep whole teams of business analysts for exactly this translation work between departments and IT. And even so, many projects fail on precisely this point. The list of abandoned software is long and the cost behind it is enormous. Most companies do not have those resources in the first place.

Why this did not work before

Even years ago there were people who could take in a company's context and build a working solution from it. It would still have failed most of the time, because you as the client could not have found the time to explain everything in detail. What was missing was not the competence, it was access to the context.

What AI changes about this

This is exactly where AI comes in. We build a large part of the context ourselves. Some of it comes from publicly available sources, just as our expert in the example already knows the concepts of “living”, “kitchen” and “appliance”.

What can only come from inside your company, the concrete workflows, systems and processes, no longer has to be laboriously prepared for us. You simply give us the data that exists anyway: calendars, notes, contracts, records, reports.

Taking all of that in would be extremely laborious and time-consuming for a single person. AI does it very efficiently. Large volumes of data are structured, filtered and related to each other. Within a few days a semantic network emerges that maps your domain. For a comparable result we would once have needed months.

And what about hallucinations?

The question of reliability is a fair one. AI can sound convincing and still deliver false information. That is why we work with validation techniques that make hallucinations extremely unlikely.

For maximum certainty there is one single, focused review session at the end. There you check and confirm the context we have built. Efficiently, and without you having to set aside weeks for it.

Why Datamantics

What sets us apart from classic software firms

IT is not an end in itself for us. We do not build systems because they are technically interesting or because the industry is busy with the next hype. If a feature creates no real value, it does not belong in the system. And if a standard product fits better than a custom build, we will tell you so.

Software has to be understood by the departments that use it. The people who work with it every day should be able to steer it, extend it and use it sensibly, without needing a provider for every small thing. A good system makes a company more independent, not more dependent.

That is why we do not want to become indispensable. Our job is to pass on knowledge, create structure and build solutions that belong to your company.

  • No off-the-shelf software

    The solution follows your processes. And if a standard product fits better, we recommend it instead of building something.

  • Fast results

    Building the context takes us days rather than months. That is why you see solid results in weeks, not after a long preliminary study.

  • Industry knowledge meets technology

    We learn your professional vocabulary and talk to your specialists as equals, not through tickets.

  • Traceable results

    Every statement the AI makes traces back to a document, a rule or a standard. Validation is part of the method, not an extra.

  • Partnership, not handover

    We work in manageable packages with clear goals. At the end, the solution, the structure and the documentation are yours.

The founder

Albert Lacambra

Behind Datamantics stands Albert Lacambra. For more than 20 years he has worked as a data analyst, software architect, IT strategist and AI pioneer, and he knows precisely how language models work. That is what the validation techniques we use to rule out hallucinations are built on.

With exactly this approach he has already made huge, organically grown software solutions obsolete at several companies. Systems that had grown for years and only ever solved the problem inadequately were replaced in a very short time, by solutions that work better, faster and more purposefully.

The method is therefore not a theory. It has been proven in practice.

Let us talk about your context.

Tell us where your systems fall short today. We will tell you honestly whether we are the right partner for it - and if not, who might be.

Book a conversation – 60 minutes