conusweb
conuswebHow to choose a custom software supplier in SK/CZ, and what to ask
Five areas to go through before signing: architecture, code ownership, GDPR, verifiable references and post-launch support. Plus the warning signs.
When choosing a custom software supplier in Slovakia or Czechia you should ask about five areas: the technical architecture of the solution, ownership of the source code and data, the approach to security and GDPR, references with verifiable results, and how the relationship works after go-live. A supplier who cannot answer clearly on architecture, code ownership or compliance probably has no experience with larger projects and is better suited to simple, standardised work.
Architecture and technical stack
Ask what stack the supplier uses and why. A serious supplier can explain the decisions, not just list technologies.
- What framework and architecture are they proposing for your specific case, and why this one?
- How will the solution scale as data volume or user numbers grow?
- If it is an AI solution, what is the architecture of the data layer: RAG, vector database, data refresh?
- Is the solution modular, or does it become one closed monolith?
Ownership of code, data and infrastructure
This is where companies most often go wrong: they sign without clarity on who actually owns the output.
- Who owns the source code once the project is finished, you or the supplier?
- Does the solution run on your infrastructure, or are you tied to the supplier's platform?
- Can the project be handed to another supplier or an internal team without losing functionality?
- Where does the data physically sit, and who has access to it?
Security and GDPR
On B2B projects, particularly in healthcare, law and finance, this is a condition rather than a bonus.
- How does the supplier ensure GDPR compliance when processing personal or sensitive data?
- Is there a data processing agreement, and is it clear who is controller and who is processor?
- How is security handled for AI components, for instance data leaking through third-party APIs? The list of what can go wrong is in the security risks of deploying an LLM.
- Is on-premise or EU-hosted deployment possible if your sector requires it?
References and verifiable results
Marketing testimonials along the lines of "great to work with" tell you very little. Ask something more specific.
- Can they show a project with measurable results, such as reduced processing time or cost savings?
- Can a reference client be contacted directly?
- What kinds of company does the supplier usually serve, and does that match your profile?
- Do they have a profile on independent platforms with verifiable reviews?
Working relationship and post-launch support
Custom software is not a one-off delivery but a long relationship. Clarify this before signing, not after.
- What happens after launch: who handles bugs, updates, scaling?
- Is there a service level agreement, and what are the response times?
- How are changes of scope handled during development?
- Is communication direct with the development team, or through several layers of intermediaries?
Warning signs
| Signal | Why it is a problem |
|---|---|
| A price far below the market | Often means thin testing, missing documentation, or hidden costs later |
| A vague answer on code ownership | Risk of dependence on a single supplier |
| Missing or only anonymous references | Real experience is hard to verify |
| No mention of GDPR unless prompted | Compliance is not a design priority |
| A fixed price without a clearly defined scope | High risk of disputes about what was included |
For comparing price levels, what an AI agent costs and custom software versus SaaS both help.
Agency versus software studio
A traditional marketing or web design agency generally optimises for speed and standardised solutions: templates, off-the-shelf CMS platforms. A software studio treats the project as an architectural problem and designs for scalability, integrations and long-term maintenance. For a simple brochure site that difference may not matter. For systems processing company data, AI automation, or solutions tied to internal processes, the architectural approach decides both stability and security.
Frequently asked questions
- Is a cheaper supplier always the worse choice?
- Not automatically, but a price far below the market average usually means a compromise in testing, documentation or security that only shows up later.
- Should I choose a local or a foreign supplier?
- A local supplier has the advantage of direct communication, familiarity with local legislation and overlapping working hours. For specific technologies a foreign specialist can be relevant. More important than geography is verifying experience with your particular type of project.
- How do I tell whether a supplier really understands AI or is just using buzzwords?
- Ask about concrete architecture: how exactly RAG works in their solution, which vector database they use, how they handle data refresh. A supplier who can only answer in generalities without technical detail probably lacks deeper implementation experience.
- How many suppliers should I approach before deciding?
- Comparing three or four suppliers with a similar focus is advisable, so you can weigh their approach to architecture, pricing and communication objectively, not just the final number.
Related articles
What a custom AI agent costs to build for a B2B company
What actually drives the price of an AI agent: scope of autonomy, integrations, RAG infrastructure, compliance, and the running costs suppliers rarely mention up front.
Read the article
How a multi-tenant vector database serves many clients at once
One AI infrastructure for dozens of clients without their data mixing: metadata filtering, the architectural decisions that matter, and the isolation tests.
Read the article
Security risks when deploying an LLM into company processes
Six risks that come not from the model but from the architecture around it: data leakage, prompt injection, access separation, agent actions, hallucination and auditing.
Read the article