No-Code, Low-Code & AI at the BFH PM Club
On June 23, 2026, we were invited to the Bern University of Applied Sciences (BFH) — as guests of the PM Club / HERMES Connect Hub for an evening dedicated to a topic that is currently on many organizations’ minds: No-Code, Low-Code, and AI.
Together with Andreas Hosang (Hosang Consulting GmbH, Board Member of the HERMES Connect Hub) and Tobias Kluge (incratec GmbH), we examined the same topic from three distinct angles: past, present, and future. Not as a panel discussion, but as three interconnected perspectives — with the goal of providing real orientation to the roughly 80 project managers in the room.
This post summarizes the key insights from the evening.
Part 1 – Past: Managing Software Instead of Building It
Andreas Hosang opened the evening with a fundamental question: How does software actually come to be — when you manage it rather than build it yourself?
The starting point was classic project reality. Traditional software projects follow a clearly structured pattern: architecture, requirements, role separation, governance, testing. The technical effort comes first; operations follow. No-Code and Low-Code shift this picture — but not in the way most people expect.
Technically simpler does not automatically mean easier to manage.
Technology makes the building part easier. But the real challenges — alignment, role clarity, maintainability — don’t go away. They shift from the technical layer to people and process. When you lead a No-Code project, you suddenly have very short build phases and very long consensus and operations phases.
Andreas broke down the role reality by project size: small initiatives are often developed and maintained by the business unit alone. As scope grows, dedicated project management becomes not just useful but critical. A larger initiative with many stakeholders, migration needs, and integration requirements can be highly complex even when the underlying technology looks simple.
Drawing on his experience across hundreds of projects — from small Power Apps to enterprise-wide SharePoint rollouts — he derived four principles:
- Slow down when it’s urgent — clarify the baseline, expectations, and complexity first
- Plan to win — be honest about team capabilities and risks
- Steer the project, strengthen the talent — budget for coaching, don’t leave junior staff unsupported
- Back and forth empties the pockets — define clear handoffs between business units and hold them
The message from Part 1 was clear: No-Code and Low-Code are not a shortcut around good project management. They represent a different project reality — with different, not fewer, demands on leadership and structure.
Part 2 – Present: When Does Low-Code Actually Pay Off?
Philippe started from a question that comes up repeatedly in practice: When should you use Low-Code — and when shouldn’t you?
The answer is not a platform recommendation. It starts with a mindset:
Not Low-Code at any cost, but the right technology for the right problem.
No-Code, Low-Code, and Pro-Code are not competing technologies. They form a spectrum. A vacation request form can perfectly well be No-Code. A global core banking platform probably cannot. The decision depends on requirements, risk, lifespan, and strategic importance — not on the label of the technology.
Used correctly, Low-Code significantly shortens development cycles, brings the business closer to the solution, and reduces the burden on IT. Used incorrectly, it creates new dependencies, uncontrolled applications, and technical debt.
Low-Code works particularly well when:
- Business logic is clear and relatively stable
- Fast iteration is important
- The business unit is the primary user or co-designer of the solution
- The use case follows standard patterns: forms, workflows, dashboards, reporting
Low-Code is less suitable for highly complex algorithms, extreme performance requirements, or very specific legacy integrations.
Philippe positioned Low-Code as the middle ground between Build and Buy: Configure — faster than building from scratch, more flexible than a rigid off-the-shelf product.
Low-Code Is Not No-Brain
First things first: No-Code and Low-Code do not mean you no longer have to think. Business process complexity and data quality do not improve simply because the technology changed.
Jumping from an unstructured Excel process directly to an AI-generated application means, in the worst case, digitizing existing chaos at higher speed. Many organizations are still on the lower rungs: processes run through Excel, Access, or email. Instead of first bringing these foundations into shape, many try to leap straight to the AI level.
No-Code and Low-Code can serve as an important intermediate step here — bringing data, processes, permissions, and business logic into a structured form. AI can then build on that foundation.
Risks That Are Often Overlooked at the Start
Low-Code reduces development effort but does not eliminate risk. A central concern is vendor lock-in: if data or business logic cannot be exported, a future migration becomes expensive or practically impossible. Add to that third-party dependencies, data protection requirements (Swiss revDSG, EU GDPR), and licensing costs that can grow surprisingly fast as usage scales.
Governance is not a brake — it is the prerequisite for Low-Code working durably inside an organization.
AI lowers the barrier to software development, but not the responsibility that comes with it.
Low-Code and AI: An Increasingly Powerful Combination
An AI agent can access a Low-Code platform via the Model Context Protocol (MCP) and create data models, configure workflows, or execute actions — within defined governance boundaries. The platform becomes the controlled execution layer between AI and operational systems.
Low-Code can do for AI what an operating system does for applications: provide a controlled, standardized environment with clear permissions, data models, and auditability. Instead of generating thousands of lines of code, the AI works with defined, validated building blocks — and the results become more stable, better documented, and easier to maintain.
Part 3 – Future: The Changing Role of the Project Manager
Tobias Kluge closed the evening with a question that preoccupies many in the field: What does AI change about the work of project managers — and what does it not?
His answer: the role shifts fundamentally — but it does not disappear.
Until now, a large part of the workload involved creating and coordinating artifacts: specifications, process descriptions, status reports. This work is increasingly moving to the machine. Expert judgment, decisions, and accountability remain with the human.
The role shifts from creating and coordinating artifacts to evaluating and taking responsibility for outcomes.
Tobias illustrated this through three theses:
Thesis 1 – Methodological knowledge becomes an on-demand resource. HERMES knowledge no longer has to be looked up in a manual — an AI agent can provide it situationally, tailored to the specific project. This changes how project managers engage with methodology.
Thesis 2 – Human value creation shifts toward review. What used to take hours — process descriptions, requirements documents — can now be produced in minutes. The value now lies in assessing and approving these outputs with expert judgment.
Thesis 3 – Properly stored information closes the loop. When tasks and documents are stored in a way that AI can directly access, manual handoffs disappear. The project manager steers and approves — specifying rather than executing.
Tobias drew a clear distinction between Vibe Coding (fast, intuitive, suitable for prototypes — risky in production) and Agentic Engineering (define the outcome, constrain the system, verify the result — governance and operational reliability included). The difference is not only technical: it is a question of responsibility.
What Remains
The evening showed that No-Code, Low-Code, and AI is too multi-layered a topic to reduce to a single statement. But one thought ran through all three contributions:
Building gets easier and faster. Leading does not. Alignment, role clarity, and maintainability become more critical — not less important.
Technology shifts where the effort goes. It does not eliminate the effort. Those who understand this can deploy Low-Code and AI with purpose — and avoid creating digitized chaos that now emerges faster than ever before.
Many thanks to the BFH, the PM Club, and the HERMES Connect Hub for the invitation and an excellent evening.
Do you have questions about the content or want to explore this topic for your organization? Get in touch with us.