2025 marked a decisive shift in how software is built.
AI stopped being experimental and became a practical, daily tool — not through hype, but through structure, tooling, and disciplined integration.
In this five-part series, we share how this transition unfolded in real-world OpenEdge projects: from the turning point itself, to tooling decisions, continuous change, evolving developer roles, and finally AI-driven modernization with the SmartComponent Library.
These articles are not predictions. They are lessons learned.
2025 as a Watershed Moment
For many in the software industry, AI had been a recurring headline topic for years.
For us, 2025 became the year where AI stopped being an experiment and started becoming an integral part of how software is actually built.
The shift wasn’t sudden — it was a progression:
- 2023: cautious experimentation (AI-driven backend access in workshops; SmartComponent Library + early ChatGPT).
- Early 2024: AI still the exception; skepticism was common, especially in the OpenEdge ecosystem.
- Mid 2025: mainstream momentum; AI dominated major conferences — but real value came from disciplined use, not hype.
This shift did not happen overnight.
Our early encounters with AI, particularly in 2023, were cautious and exploratory. We experimented with AI-driven access to backends during workshops, such as those conducted at the PUG Challenge conference, where we exposed standard RESTful endpoints on top of a Sports2000 database using the SmartComponent Library.
Combined with early versions of ChatGPT, the results were already impressive — but still experimental.
At the beginning of 2024, AI usage in mainstream software development was still the exception rather than the rule.
In discussions with customers, we frequently encountered skepticism, especially in the OpenEdge ecosystem. Claims circulated that AI would massively favor C# or Java developers, leaving ABL developers and development teams behind in terms of efficiency.
Reality, however, told a different story.
By mid-2025, the narrative changed decisively.
At the WE ARE DEVELOPERS conference in Berlin, AI dominated nearly every keynote and session. Even high-profile demos by the likes of GitHub CEO Thomas Dohmke revealed that AI was powerful — but far from magical.
Key takeaway: Productivity gains come not from hype, but from disciplined integration.
When AI Moved Closer to the Code
The real breakthrough came when AI moved closer to the codebase, architecture, and domain knowledge.
With the release of the MCP Connector for OpenEdge in early 2026 — which we were able to evaluate in close cooperation with Progress Software — AI agents finally gained structured access to:
Structured access changed the quality of AI output:
- ABL references
- Framework knowledge
- Architectural context
At this point, something fundamental changed.
AI stopped guessing. It started understanding.
Why This Matters
2025 demonstrated that AI is not about replacing technologies or developers.
It is about augmenting structured ecosystems - and OpenEdge, combined with strong frameworks, is exceptionally well positioned for this future.
Outlook to Part 2
In the next part, we take a deep dive into the tools that enabled this transformation and explain why development tooling has become a strategic decision in the age of AI.