AI SOFTWARE
DEVELOPMENT

Depends on scope, but our default is a dedicated team that takes full ownership of the build, or a team extension if you already have engineers and just need to fill an AI-specific gap. Either way, it’s senior engineers only, and the team stays on your project exclusively, no split attention across other clients.

Either works. If you already know what you want built, we can move straight into planning. If you’re still working out the shape of it, we usually start with a short discovery phase to nail down scope before committing to a full build, so you’re not paying for engineering time on something that’s still changing.

Not every problem starts with AI. Sometimes the right fix is a solid, custom-built system, with AI added once the foundation can actually support it.

Quite a lot, depending on where you’re starting from:

• AI opportunity assessment, figuring out where AI is actually worth your time before you spend anything on development
• AI integration, connecting AI to systems you already run: ERPs, CRMs, internal tools
• Custom AI development, from LLM-powered features like copilots and intelligent search to recommendation engines and document processing
• AI agent development, agents that handle full workflows end to end, not just answer questions
• Real-time AI features like voice assistants and chatbots
• Legacy modernization, when the right first move is a solid custom system rather than AI
• Private AI on Spark, our in-house infrastructure, for clients in regulated industries who can’t send data to a public API

Honestly, it’s hard to give a real number without knowing your specific situation, scope changes the price more than almost anything else. As a rough ballpark: a focused AI feature or agentic workflow engagement usually starts around €80k to €150k. Larger, full production builds or legacy modernization projects tied to AI tend to run €250k and up. Smaller discovery or readiness work, figuring out what’s worth building before you commit to anything, is typically €15k to €50k. The best way to get a number that actually means something is the call itself, that’s where we scope your specific idea and give you a ballpark you can actually plan around.

Yes. For regulated environments, we run models locally on Spark, our in-house AI workstation, so nothing is sent to external providers or public APIs during development, testing, or integration.

No. We build across OpenAI, Anthropic, Gemini, and open-source models, so your architecture doesn’t depend on one company’s pricing or roadmap. If a model provider changes course, you’re not stuck rebuilding from scratch.

Hard to say precisely without knowing your scope, that’s what moves the number most. As a rough ballpark, a scoped agentic workflow or AI feature engagement typically runs €150k to €300k. Larger or longer-running programs, especially ones that include modernization work, tend to land at €500k+. The real number comes out of the first conversation, once we understand what you’re actually building.

We check before we commit to anything. If the data or infrastructure needs work first, we scope that plainly and either do it or hand you a clear spec.

We build every system with security in mind from day one, not bolted on for a procurement review. We’re ISO/IEC 27001 certified, and depending on your industry, we build toward HIPAA, GDPR, or SOC 2 requirements as well. That includes secure coding practices, access control, and audit trails throughout, not just at handoff.