Budapest, Hungary
AI consulting in Hungary.
We have been building software from Budapest since 2021, and AI work is now most of what clients ask us to scope. As AI consultants our first job is conversational AI, automation and AI features inside products people already use, assessed against your actual processes rather than against a demo.

Corporate LiveWire named us AI Solutions Company of the Year in 2023. Our other awards are for web, WordPress, React and SEO, and we do not dress those up as AI ones.
What we are asked for
Four jobs, and one of them is saying no.
Where AI is worth using, and where it is not
The first output is usually a shorter list than the one you arrived with. Most processes that look like AI problems are reporting problems, and we would rather tell you that in week one than bill you for a model that automates a bad process faster.
AI consultingConversational AI that can hand over
A chatbot or voice assistant that answers from your own documentation, and reaches a person with the transcript attached when it cannot finish. A bot with no exit is a customer service problem wearing a technology costume.
Custom chatbotsAutomation of the work nobody should be doing
The handoffs between systems that a person currently does by copying between tabs. We size it against your real volumes before you commit, not after your first bill.
AI automationPrototypes finished properly
If something already exists out of Lovable, v0, Bolt or Cursor, that is a specific job with a specific shape: server-side authorisation, a real data model, tests on the paths that cost money if they break.
Prototype to production
Why a team here
AI scoping is a conversation, and conversations have a timezone.
Central European Time
A question at ten in the morning gets an answer before lunch, not overnight. For a discovery phase that runs on conversations rather than tickets, that is most of the value.
English or Hungarian
Briefs, workshops and calls in either. The people you talk to are here; the engineering team is in India, and we tell you that at the start rather than letting you work it out from an invoice.
A team you can meet
In person, in Budapest, before you commit budget to something you cannot yet see. AI scoping in particular goes badly over email.
Your data stays yours
We will tell you what leaves your systems, where it goes and what it is used for, before anything is connected. If the honest answer is that a use case needs data you cannot send to a third-party model, that is a finding, not an obstacle to talk around.
How it runs
Five stages, and you can stop at any of them.
Stage 1
A conversation about the process, not the model
What the work is now, who does it, how often, and what going wrong costs. No tooling is chosen at this stage.
Stage 2
A written assessment
Which use cases are worth doing, in what order, with an estimate against each and an explicit list of what we recommend not doing. Fixed price, and yours whether or not we build it.
Stage 3
One narrow thing, in production
The smallest use case that produces a real result, running against real data with a person still in the loop. This is where assumptions meet reality cheaply.
Stage 4
Measure it against the old way
Time, error rate, cost per item. If it did not beat the process it replaced, that is the finding and we say so.
Stage 5
Widen it, or stop
Extend to the next use case, hand it to your team with documentation, or stop. All three are acceptable outcomes and the second is common.
Shipped
AI work that is in production, not in a deck.
Questions
The things people ask first.
What does AI consulting cost in Hungary?
Are you actually in Hungary?
Do you have an AI track record or just a web one?
We are not sure AI is the answer. Is that a wasted conversation?
Are you an AI consultancy or a development agency?
Will this lock us into one model vendor?
Can our own team run it afterwards?
Tell us the process. We will tell you whether AI is the answer.
Book an assessment ↗