01.1
The group worked eighteen years in the sector before incorporating the parent.
Eight years of research before the holding company was incorporated
Eighteen years inside the automotive industry and one difficulty that repeated across every one of its areas. Finding a single piece of information meant going through a dozen places that do not talk to each other. Everything else comes from there.
The problem was visible before the answer existed
Eight years from the question to the product. For most of that time nothing visible was built, and that is the part that explains why what exists today is built the way it is.
What shows from outside
Built inside
Usable outside
Before 2018
2018 · 2019
2021 · 2022
2024
2024 · 2026
2026
The sector, from the inside
A warehouse does not solve the problem
From the network to the platforms
The holding company
The long stretch
The current year
Seven of the eight years produced nothing anyone could use
The six stretches of the trajectory. It orders and does not measure, because the width of each one is not proportional to how long it lasted.
A group that starts with the company and looks for the problem afterwards arrives sooner at having a business and later at having something to sell.
01 / 06
Before 2018
The sector, from the inside
Bilal Hamida, the group's founder, spends eighteen years inside the automotive industry and across several of its areas. The same difficulty repeats in all of them. The information of the sector exists and lives split between whoever generates it and whoever needs it, and those are almost never the same party. The history of a single vehicle passes through the manufacturer, through the workshop that maintains it, through the roadworthiness test that inspects it and through the insurer that covers it, and none of the four ever sees the complete file.
02 / 06
2018 · 2019
A warehouse does not solve the problem
Gathering the information of the sector in one place leaves it just as unusable, because what arrives carries more noise than signal. The same part is named in five different ways depending on the manufacturer. One single fault enters as a diagnostic code at the workshop and as free text in the customer's claim. Something is needed that extracts, filters and cross-checks before answering, and at that point artificial intelligence stops being optional. That is where the work on the Calaa neural network begins. How it is designed, how it absorbs what arrives in twenty different formats and how it is made efficient at a scale that does not fit in a single server.
03 / 06
2021 · 2022
From the network to the platforms
This is where the architecture of the group is drawn and which platforms are needed to hold it up. The question moves from whether it can be built to where it is used. A workshop queries from the lift with its hands full, and a manufacturer from a data room with twenty columns in front of it, and those two things do not fit on the same screen. What surface each one needs is decided here, what data it asks for and what has to be settled before writing the first line of interface. It is also the stretch where the data model is fixed, the only part of the project that cannot be rebuilt later without losing the history.
04 / 06
2024
The holding company
Calaa Capital is incorporated. The work stops being research with a product at the end and becomes a group with a structure, with a single body that decides where capital goes, what gets built and by which criteria the result is judged. It is also the year the Cal-1 application is structured, and how it divides the work with the other pieces. Codex holds the record of the industry, Calaa interprets it and answers, and Calaa ID identifies whoever is asking, because a workshop and the owner of a vehicle cannot see the same part of the same file.
05 / 06
2024 · 2026
The long stretch
Training and extension of the intelligence layer, development of the applications and construction of the hardware that measures the vehicle. The three products are rebuilt and refined internally several times along the stretch. It is the part that does not show from outside and the one that lasts longest, because every piece depends on the previous one working for real and not in a demonstration. The stretch ends in 2026 with Cal-1, Calaa and Codex ready.
06 / 06
2026
The current year
Finish and go out. The group's platforms launch in October and in November. What has been built over the previous eight years is placed in front of whoever is going to use it, which is the only place where it can be checked whether it was well conceived.
Six years of research before the holding company was incorporated. A group that starts with the company and looks for the problem afterwards arrives sooner at having a business and later at having something to sell.
How far ahead it looks
Three horizons, and none of them is written in the present tense. What follows is what would have to be true for the work of the previous eight years to have been worth it.
01 / 03
2026 · 2028
That the data exists
The data exists when the four pieces mature at the same time. Cal-1 leaves its first version behind and becomes the workshop's daily tool. Calaa extends its pre-training with a larger corpus and with the concrete tasks of the trade. Codex widens the record to more makes, more markets and more years of history. Out of that work comes Cal-1 Pro+, the version for whoever operates a fleet or a workshop at scale. None of this works without continuous measurement on real vehicles and in more than one country, and it is the only one of the three horizons that depends on the group alone.
The four pieces Cal-1CalaaCodexCal-1 Pro+02 / 03
2028 · 2032
That decisions are made with it
In this stretch the circuit closes. Cal-1 Link measures the vehicle from the dashboard and KRN-01 measures it from orbit, Calaa interprets what arrives from both and Codex keeps it, and every query returns to the record what it confirms and what it disproves. Automotive information stops being gathered and starts feeding itself. What the circuit does not close is the decision. A dealer, a workshop or a fleet changing source depends on someone outside changing a habit, and a sector with tight margins changes slowly.
The circuit Cal-1 LinkKRN-01CalaaCodexThe decision03 / 03
2032 · 2035
That the sector takes it for granted
With the whole core in service, being the world reference of the sector stops being an ambition and becomes something that can be checked. A manufacturer deciding a recall, an insurer pricing a risk and a buyer looking at a history all query the same record and get the same answer. Automotive information stops being searched for in a dozen places, in the same way that nobody today looks up an address in three different maps. That is the end of the problem that set all of this in motion, and the only result that would make it unnecessary to keep explaining it.
The same record ManufacturerInsurerBuyerOne single record
The distance between data that exists and data someone decides with is measured in years, not in versions.


