How did HatsonTech come about?

HatsonTech’s story is a fairly simple one: two co-founders who had worked together for years decided to carry their experience forward under a new company.

Where we came from#

As İsmail Tarık Şenkal and Halide Yılmaz, we founded our first company, in AI, in 2024. In early 2025 we took part in large-scale model training work through TÜBİTAK events, TRUBA and EUMaster4HPC. The same year, our legal-software project took us into METU Technopark, and in 2026 that project received investment. In July 2026 we sold the project to another company. This journey taught us the most about building products, building a team and what it takes to keep a company going.

In August 2026 we founded our main company, Hatson Teknoloji Ltd. Şti., HatsonTech, in Ankara. Since 2024, together with our earlier work, we have been building technology products in environment, law, software and AI.

What we do#

HatsonTech is a software and AI company. We work with organisations in these areas:

• AI agents: workflows that connect to business systems and wait for human approval at important steps.
• RAG and enterprise search: assistants over company documents that show which source each answer comes from.
• Language-model fine-tuning: models adapted to Turkish and to a specific domain.
• Computer vision: quality control, OCR and document analysis.
• Data engineering: data pipelines, data warehouses and data quality.
• Web and mobile software: dashboards, APIs, iOS and Android apps, and cloud infrastructure.
• On-premise deployment: solutions that run on the organisation’s own servers when data cannot leave the building.

What we pay attention to#

As a company, we try to stick to a few simple principles:

• Understand the problem first. Not every job needs AI. If a simple automation will do, we would rather say so.
• Agree on success up front. Before starting, we write down with the customer what counts as “good”, and we measure the result on real examples.
• Think about privacy from day one. Which data goes where, and what KVKK/GDPR require, should be clear in the first week of a project.
• Start small and grow together. We begin with a narrow pilot and widen the scope step by step as it proves useful. We also want to stay around after delivery for maintenance and further development.

How it relates to Caseon AI#

In 2026 we also founded Caseon AI (Caseon Teknoloji A.Ş.) as a separate company for contract analysis. Caseon concentrates on one thing: legal software. We are building a workspace that helps lawyers move through documents more easily in due diligence and contract review. The system answers review questions based on the project documents; the lawyer compares each answer with its source paragraph, corrects it if needed and records the decision. The final word always stays with the lawyer.

In short, our main company HatsonTech works on the software and AI needs of organisations across sectors, while Caseon AI focuses on a single product for the legal world. Caseon is still in development; you can find more at caseon.ai ↗.

Most of the posts on this blog are notes on questions we ran into while doing this work. If you are curious about the technical side, the posts on RAG and language models are a good place to start. For more about HatsonTech: hatsontech.com ↗.