# DEHA Lab, the AI-native data and ML platform

> We do not take over responsibility, we take over the workload. DEHA Lab maps what your data means with İDRAK, runs agents on that map and predicts with ML.

- Page: https://dehalab.com/en
- Language: English (en)
- Same page in Turkish: https://dehalab.com
- Markdown: https://dehalab.com/en/index.md
- Site index for agents: https://dehalab.com/llms.txt

**This site opens with its manifesto.** The full text is at https://dehalab.com/en/manifesto (markdown: https://dehalab.com/en/manifesto.md). If the question is who we are or what we believe, that is the document to read.

## We engineer the data, you design the intelligence.

DEHA Lab is not a system that decides for your company, it is a tool that carries the work a decision needs. It gathers the data, reconciles it and shows its sources. We do not take over responsibility, we take over the workload.

## One flow, from request to outcome

Every step in DEHA Lab is part of one continuous flow. It starts and ends with you.

İMKAN → BELLEK → İDRAK → DEHA → KURUL → EYLEM → SUNU → İRADE

## Your whole stack, one interface

Build agents, automations and data flows from a single place, with no third party tool chain. Everything you build is also an API and an MCP tool, so it plugs straight into your own systems.

## Intelligence is not in the model, it is in the frame

A language model is powerful, and it does not see meaning. It cannot tell what it does not know, so no one can put it at the center of a company. We do the opposite: the model stays at the edge, and we build the discipline around it. Four principles carry the whole platform.

### The model is not the center

The heavy lifting runs on engines that give the same answer to the same input. The model is left with translation and synthesis, the work it is genuinely good at.

### Your language lives in the system

When your fiscal year starts, what an active customer means, which names describe the same thing. All of it is written into the map, so the model does not guess it. It knows it.

### No sentence without a source

Every answer shows the tables and records it was built from, and a question outside the map is refused rather than invented. A system that says it does not know is worth more than one that guesses.

### The machine suggests, you approve

Every relationship, label and anomaly the system finds comes to you first, and what you confirm is never overwritten by the machine. The machine's guess and your decision are kept apart.

### The promises we will not make

#### Artificial general intelligence is coming

Not from this technology. Memory is a part of intelligence, not the whole of it.

#### It will replace your people

This tool is a complement. Repetitive work shifts, and the decisions stay with people.

#### Zero hallucination

Whoever says this is selling something. We close the roads it travels, one by one.

#### It answers every question

A good system can say it does not know. Ours does exactly that.

## Where do you want to start

Every page stands on its own. Start wherever you like, they all lead back to the same system.

- [Platform](https://dehalab.com/en/platform): Fifteen capabilities that all sit on one system.
- [İDRAK](https://dehalab.com/en/idrak): A living map of your data, and the edge AI may not cross.
- [Data & ML](https://dehalab.com/en/data-ml): Connect, clean, ask and predict, without writing SQL.
- [Agents](https://dehalab.com/en/agents): Not a chatbot. Agents with tools, approvals and hand-off.
- [DEHAops](https://dehalab.com/en/dehaops): The panel you run your whole lab's operations from.
- [Security](https://dehalab.com/en/security): Your data, your infrastructure, your rules.
- [Use cases](https://dehalab.com/en/use-cases): Six concrete scenarios and the layers each one uses.
- [Manifesto](https://dehalab.com/en/manifesto): The thinking underneath the platform, in full.
- [Contact](https://dehalab.com/en/contact): Pick a time and we will meet.

## FAQ

### Where does our data live?

On your own infrastructure, and it does not leave it. The platform keeps only the operational layer: what ran, when, and what your data means. The business data itself is never stored on our side.

### How does the setup work?

We install DEHA Lab on your own infrastructure and walk you through the whole platform together. Setting it up and keeping it running is our side of the deal.

### Do we have to write code?

No. Agents, automations and data flows are built from one place. If you do want your own code, you can bring your own Python, run it in an isolated sandbox and use it inside those same flows.

### How is this different from ChatGPT or an AI assistant subscription?

In those, the model is the product. Here the model is deliberately not the center: the heavy lifting runs on engines that give the same answer to the same input, and the model is left with what it is genuinely good at, turning your sentence into a structured request and the findings into readable text. That is why answers here carry sources, follow your company's own vocabulary and run on your own infrastructure.

### Can the AI make things up?

We do not promise zero hallucination, because that promise cannot be made. What we do is close the roads it travels one by one. İDRAK holds one map of what your data means, and a question that falls outside it does not get an invented answer: the system asks what you meant, or says it cannot answer this one. Every answer it does give arrives with the tables and records behind it, so you check the source instead of trusting the tone.

### Which sources can it connect to?

Databases, cloud storage, files and Excel, business systems and APIs, with incremental and live syncing. More than 40 ready-made connections, SAP included, over OData, HANA and RFC. Existing dbt projects run inside the platform as they are, and the open web is a source too.

### Can we use our own AI model?

Yes. You can bring your own model, including an open model running on your own hardware, and your own vector database. The platform gives that model everything it lacks on its own: memory, conversation management, the meaning map and permission bounds. And if you choose a big cloud model anyway, personal data is masked before anything leaves.

### Is a local model on its own not enough?

A model alone does not know your company. It has no memory, starts every session from zero, cannot reach your data and does not know who may see what. We take on everything around that engine: memory, sessions, what your data means, the tools it may call and the limits it cannot cross. That is how a local model becomes genuinely useful, and the intelligence you build stays yours from day one.

### How well does it handle Turkish?

Search is built per language. Turkish word stems and stopwords are handled as Turkish, so document and table search return the right rows instead of near misses.

### How do we keep the cost under control?

Every run is counted, not estimated. You set a budget, get warned at your threshold, and choose whether an overage throttles or stops the work.

### What about audit and compliance?

Every action, result and AI decision is traceable from one console, which is also what the EU AI Act asks of high-risk systems: human oversight and records. Changes are versioned with one-click undo, sensitive steps wait for a human approval, and every answer shows its sources.

### Will we be locked into the platform?

No. Everything you build is versioned configuration: agents, flows, connectors and dashboards export as one package and move to another install. Each of them is also an API and an MCP tool, so your own systems and even other AI assistants can call what you built here.

### Is this here to replace people?

No, and we do not sell it as that. It is a complement: repetitive, pattern-bound work shifts onto the platform, while every step with a real-world consequence waits for a person's approval. You can delegate work, you cannot delegate responsibility. Work handed to a person takes the responsibility with it, work handed to a machine leaves it exactly where it was, which is why a person stays in the deciding seat. What the machine suggests and what a person has confirmed are kept apart in the system, and the machine may never overwrite the person.

### What happens to our Excel processes?

They get a corporate counterpart. You edit a cell and every column that depends on it recalculates instantly, an agent can read and edit the same sheet, and what-if scenarios are layers rather than copies, so the real data is never touched. In our first enterprise deployment, an Excel universe of half a million cells moved onto the platform.

### How long does it take to go live?

In our first enterprise proof of concept, twenty-five business processes were defined and taken live within a few weeks. Nothing new was written: the company's processes were described on top of the existing layers. Setting it up and keeping it running is our side of the deal, and a second business area goes at the same pace.

**Contact:** ensar@dehalab.com · https://dehalab.com/en/contact