Just a few years ago, building data architectures was largely manual craftsmanship: fine-tuning SQL queries, debugging Python scripts, and constructing heavy ETL pipelines. Then generative AI arrived — and the playing field changed permanently.
I have been following the data industry, software development, and consulting for over 25 years, since the turn of the millennium. I have seen technologies rise and fade. Yet I can confidently say that the transformation we are witnessing now is the most significant shift since the move to the cloud began in the 2010s.
While the cloud changed where data resides and where processing happens, AI is changing how we work with data on a daily basis.
This does not mean data professionals are disappearing. It means the emphasis of the role is shifting: from hands-on coders to architects, curators, and quality controllers.
Software giants at the forefront of AI
Major data platforms and integration tools have not stood still. They have embedded AI directly into their core capabilities automating a lot of work.
Some examples:
| Platform | Examples of AI Solutions |
| Snowflake (Cortex) | Brings LLM models directly to the data, enabling intelligent analysis without moving data outside secure environments. |
| Databricks (Agent Bricks) | Provides tools to build autonomous agents capable of retrieving information and executing tasks based on enterprise data. |
| Microsoft Fabric | Copilot guides users across the entire data flow, automating code generation, documentation, and reporting. |
| SAP (Joule) | Simplifies navigation within complex ERP data structures and translates business processes into faster, clearer insights. |
“Vibe” and Agent Coding require true seniority
One of the most interesting emerging trends is vibe coding. In this model, the expert does not necessarily write every line of code themselves. Instead, they describe the desired outcome in natural language – and AI handles much of the technical implementation. The focus shifts from syntax to intention.
Another rapidly emerging concept is agent coding.
Agent coding refers to a model in which AI agents operate partially or fully autonomously: they retrieve information, process data, and execute defined tasks within an enterprise data environment based on predefined objectives. In this model, the role of the expert shifts from hands-on implementation to supervision, orchestration, and governance.
However, there is an important risk to consider: if you do not understand what high-quality code and sustainable architecture are supposed to feel like, vibe coding can lead to solutions that are difficult to maintain and evolve over time. AI is an excellent assistant, but it still requires human guidance.
This is where seniority and experience become essential. Experienced professionals quickly recognize when an AI-generated solution is technically impressive but architecturally weak, insecure, or hard to scale. Paradoxically, the easier code generation becomes, the more important deep expertise is.
As organizations move toward broader agent-based environments, where multiple AI agents operate semi-autonomously, the ability to evaluate, orchestrate, and govern these systems becomes increasingly important.
At Etlia, we are at the core of this development
This technological shift is unfolding faster than any previous transformation. That is why at Etlia, we invest heavily in continuous learning and active market monitoring, in ways that directly benefit our clients.
- Active monitoring: we continuously analyze announcements from market leaders. We do not rely on marketing materials alone – we test new features in practice to understand their real value. Credit is also due to technology vendors who keep partners well informed about new AI capabilities.
- Validation of expertise: we use AI to accelerate routine tasks, but senior-level expertise always ensures that the end result is production-ready, secure, and maintainable.
- Knowledge sharing: we actively share insights within our team so that the latest understanding immediately translates into value for our clients.
For us, it is a matter of professional pride that our clients receive up-to-date strategic guidance. We do not build yesterday’s solutions for today’s problems.
Is traditional expertise still needed?
Although AI handles many routine tasks, a deep understanding of data structures and integrations is more critical than ever. The role of the expert has shifted from execution to oversight, architecture, and orchestration. One must know precisely what is being built in order to guide AI in the right direction.
Data engineering is evolving from technical execution toward a more strategic function. As native AI tools within platforms handle heavy lifting, more space is created for what truly matters: turning data into real business value.
-Petri Räsänen
If you would like to explore how to navigate this transformation and adopt AI tools in a controlled and sustainable way, our experts at Etlia are ready to support you.
