How it works
Every service we build, and every service built on our technology, runs on the same core technology and the same six layers: concept, code, intelligence, rendering, verification and reality.
The servers, storage and networks our services run on.
Collecting, storing and updating large datasets from named sources.
Structuring raw data into models of real conditions.
The programs that run the calculations behind each model.
Secure access to our data and models from other systems, with managed keys and usage.
Processes that collect data, run models and refresh results without manual steps.
Custom code and applications built on our engines.
The maps, tables, dashboards and 3D views people use to work with the results.
Six layers
The five layers of every model of Octa Models, which is built on our technology, and a sixth that holds them all to the truth: the reality layer. Concept, code, rendering and reality are in every service; the intelligence and verification layers are the ones the tiers add.
The Octa Methodology: what to ask, and in what order.
01code layerMathematics and relationshipsReal data and established science. Where there is no data, we show a gap rather than a guess.
02intelligence layerEngine and AIAI that reasons within limits: inside our models, traced to the sources behind them.
03rendering layerData and presentationThe maps, tables, dashboards and 3D views people use to work with the results.
04verification layerHuman intelligencePeople who check the results against evidence and the real world.
05reality layerWhat is realOur best way to determine what is real.
Built on the Octa Methodology.
Our work is built on the Octa Methodology. Its eight aspects define the domains we model and analyse, and its eight sections define the core services we build our intelligence models for.
In every serviceBuilt on real data and established science.
Our models are built from real data and established methods in physics, mathematics, ecology, biology and the indicators used to track economies, governments, societies and technologies. We use each country's official data first, open global data where there is none, and show a gap where no data exists. Every figure names its source.
Each calculation follows a published method and cites its reference.
In every serviceStructured environments for AI inference.
AI is an optional second layer, connected through APIs. It works inside the models we build, which serve as its source of truth, so every answer can be traced to the sources of data behind it. This limits AI to narrower applications and reduces the risk of hallucination.
Tier 2 and Tier 3The maps, tables, dashboards and 3D views people use to work with the results, and secure access to our data and models from other systems through our APIs, with managed keys and usage.
In every serviceSpecialists check what matters.
Where accuracy matters most, specialists check results against evidence: fieldwork, laboratory tests, expert review or other methods. Verified results are labelled as verified. When a check on the ground disagrees with the model, the ground takes precedence and we correct the model.
Tier 3 · quoted before work beginsOur best way to determine what is real.
Seeing, knowing and taking action are only as good as the truth beneath them, and truth is hard to develop. Data is incomplete, sources disagree, models drift, and AI can invent.
The reality layer is how we develop it. Every value carries its source, and every gap is shown as a gap. AI answers only from our models. People check what matters against the real world, and when the ground and the model disagree, the ground wins and the model is corrected. What holds through all five layers beneath it is what we treat as real.
In every serviceSee. Know. Act.
Seeing, knowing and taking action must be based on truth, and truth is hard to develop. Data is incomplete, sources disagree, models drift and AI can invent.
Our six layers are how we develop it, one on another, out to the reality layer: our best way to determine what is real.
Seeing starts with what is there. Every value names its source, and where there is no data we show a gap, never a guess.
Knowing is understanding what it means: published methods, and AI that answers only from our models, so every answer can be traced.
Acting is the test. People check what matters on the ground, and the ground has the last word: the model is corrected, and every service built on it benefits.
Our hardware, data, models and engines are our own. Outside services, such as AI models or mapping imagery, connect through their APIs where they are useful, and nothing at the core depends on them. Octa HQ's Fruits, the independent companies it incubates, use this technology directly.
See OperationsWhat we will not build
We do not build or support systems that monitor or control people. This condition applies to every service and every access key we issue.
We price every service the same way, and so do the Fruits: the operational cost of the work, meaning the compute and API calls it uses, plus a 20% fee on those costs. The fee is charged on usage, so it continues for as long as a service runs. Human verification and custom development are quoted before work begins.
We adapt the eight sections of the Octa Methodology into our services.
| Service | Status |
|---|---|
| Objective and Goals Aligner | Available |
| Site selection | Available |
| Data collection | Available |
| Impact assessment | In development |
| Risk analysis | In development |
| Trend forecasting | In development |
| Concept development | Planned |
| Work specification | Planned |
Tell us what you need to know and we will tell you which services apply, what we would build and what it would cost.
For now, we take on only projects aligned with the Octa Methodology and the work of the Octa HQ incubator.