.svg){kind=small}
Rólunk
Digitalizáljuk vállalati folyamatait! Együtt! Hatékonyan!
Tudjon meg többetAdatkezelési tájékoztatóImpresszumBiztonság és megfelelőségElérhetőségeink
Cím
1095 Budapest, Soroksári út 48. (Hungária Malomudvar, 8. épület 2. emelet)
Situations you will recognise
Shadow AI
Your staff already use ChatGPT — with company documents, without oversight. Banning it does not work; leaving it unregulated is a risk. You need an internal alternative that is better than what they are using in secret.
Legal said no to the cloud
The AI project is ready, everyone is enthusiastic, and then the data protection officer asks the question: where does the data go? And the project stops. For us this is where it starts: on-premise and hybrid architectures take AI into places cloud vendors walk away from.
The knowledge exists, nobody can find it
Thousands of documents, three systems, and know-how in the heads of colleagues approaching retirement. A permission-aware knowledge base agent answers in minutes from material that takes half a day to track down today.
The back office is people scrolling
Quote requests, invoices, email triage, reconciling data across systems. These processes are candidates for agent automation — with people deciding and AI doing the work.
What we build
- Internal knowledge base assistant — answers from your own documents, with source references, respecting access rights
- Document processing agents — extracting and structuring contracts, invoices and technical documentation
- Quote automation — from the incoming request to a calculated draft quote
- Customer service AI — from web chat to email triage, connected to your existing systems
- ERP and legacy integration — on Java/Angular foundations; including where no-code tools run out of road
We deliver this two ways: built on our own Athene AI platform (faster, with predictable licence costs), or as fully custom development if the process calls for it. The two do not compete: the assessment decides which is worth it for you.
Here the project does not stop at compliance
We have run AI agents where the data never travelled a single metre
Open-weight language models (Llama, Mistral) on the company's own GPU server, in a private cloud or in a hybrid setup — in production, not in a lab. Few vendors on the Hungarian market can demonstrate this today; most "enterprise AI" is in fact a front end on a US API.
In your own language
We tune the models to Hungarian enterprise language — the agent does not answer your colleagues in translation-flavoured Hungarian.
We produce the EU AI Act documentation
Risk classification, transparency elements, data processing agreement — part of the deployment, not an extra charge. Details on the Security and compliance page.
Senior team, reachable leadership
An average of 10+ years of development experience, and your project does not disappear into an account hierarchy: you talk to the decision-maker, not to a ticketing system.
How we work
1 · Assessment (2 weeks)
Processes, data assets, infrastructure, compliance constraints. It ends with a written proposal: what is worth handing to agents, on what architecture, at what cost.
2 · Proof of concept (4–6 weeks)
A working system on real data, with success criteria agreed in advance and a go/no-go decision point. The usual pilot on the market takes 90 days — we deliver in half that time.
3 · Rollout and operations
Gradual expansion, training, operations backed by an SLA. The system stays with you — together with the source code and the documentation.
Our clients trust us with their processes
Legal · Privacy Policy · Legal Notice
