AI & automation in Malta

Put AI and automation to work on your real systems and data, safely: from connecting an AI assistant to your company tools to automating the processes that eat your team's time.

What I can do for you

  • AI assistants that know your businessConnect Claude or GPT to your documents, databases and tools through the Model Context Protocol (MCP), so answers come from your own data.
  • Workflow automationAutomate repetitive work with n8n: approvals, data entry, notifications and syncing between the systems you already use.
  • AI features in your productDesign and build AI copilots and assistants inside your own software.
  • AI readiness & governanceFind where AI genuinely pays off, what data it can safely see, and the controls you need before rolling it out.

Why work with me

  • I've done this inside a large SaaS company. At NetRefer I built an MCP aggregator that connected every company system to Claude Enterprise. Early results compressed development cycles from months to days.
  • I build the tools other teams use. I founded ApexMCP, a gateway that connects databases, APIs and SaaS tools to any AI agent through one endpoint, with encrypted credentials, scoped keys and full audit logs.
  • I've shipped automation in production. I introduced n8n at NetRefer, both for internal workflows and inside the client-facing product, and designed several iterations of its AI CoPilot.
  • I take the risks seriously. I advise leadership on AI governance and data access, not just the technology.

How a project works

  1. DiscoveryWe map the processes and data where AI or automation could help most.
  2. ProposalA short plan with the first use case, what it needs access to, and a price.
  3. PilotA working version on real data, so you can judge the results before scaling.
  4. Roll-outProduction set-up with access controls, monitoring and handover to your team.

Questions

Is my company data safe with an AI assistant?

I design integrations so the AI only sees what it needs: read-only access where possible, credentials kept in a secure vault, separate keys per team, and a log of every action. We agree what data is in scope before anything is connected.

Do I need a big budget to start?

No. Most projects start with one well-chosen process or use case, proven on real data, before anything larger.

Which AI models do you work with?

Mainly Claude (Anthropic) and GPT (OpenAI), connected through MCP, which also works with tools such as Cursor and Windsurf.

Have a process you'd like to automate, or an AI idea to test? Tell me about it.