30% Saas vs Software Shift Isn't What You Thought
— 6 min read
The 30% SaaS vs software shift isn’t about moving a third of apps to the cloud; it’s about AI auto-code platforms replacing many SaaS plugins, cutting coding time by up to 90%.
The Real Meaning of the 30% Shift
When the headline "30% SaaS vs Software shift" first hit the tech wires, most of us pictured a tidy migration chart - 30 per cent of on-prem tools disappearing overnight. In truth, the shift is far messier, driven by a blend of economic pressure, talent shortages and, increasingly, AI-driven code generation. I was talking to a publican in Galway last month, and even he could see the change: his bar’s booking system was built with a handful of SaaS add-ons, but a new AI tool promised to rewrite the whole stack in days, not months.
Sam Altman’s recent proclamation that SaaS is "dead" (OpenTools) isn’t a death sentence for the model; it’s a call to evolve. Companies are still buying SaaS, but they’re also looking for ways to cut the recurring licence fees that add up quickly. The 30 per cent figure, cited in several analyst decks, actually reflects the proportion of development effort that can now be automated - not the share of applications that will disappear.
According to the latest CSO data, Irish firms that adopted AI-assisted development reported a 12 per cent uplift in delivery speed, echoing global trends (Netguru). This isn’t a marginal tweak; it’s a structural change that reshapes how we think about "software" versus "service". The old line - SaaS on the left, on-premise on the right - is blurring, with AI code generators sitting somewhere in the middle, offering a hybrid that feels more like a new category altogether.
Here’s the thing about the 30 per cent myth: it masks a deeper question - are we replacing software with services, or are we simply automating the software-building process? The answer determines where the money goes, where the jobs are, and how we, as Irish developers, need to upskill.
Key Takeaways
- 30% shift refers to automatable development effort, not app migration.
- AI auto-code tools can cut coding time by up to 90%.
- SaaS plugins are being superseded by AI-generated code.
- Irish firms see faster delivery and lower licence costs.
- Future skillsets will focus on prompting and AI oversight.
How AI Auto-Code Engines Are Cutting Coding Time
Imagine cutting your coding time by 90 per cent - that’s the promise of the newest generation of autonomous code platforms. In my own work with a fintech start-up in Dublin, we trialled Legato’s AI-driven “vibe” builder (Legato recent funding). Within two weeks the team produced a prototype of a compliance dashboard that would have taken months using traditional SaaS plugins.
The technology sits on three pillars: large language models trained on billions of lines of code, prompt-engineering interfaces that let non-engineers describe functionality in plain English, and continuous integration pipelines that automatically test and deploy the generated code. The result is a loop where a product owner can type "Create a secure user login with two-factor authentication" and watch a full stack - front-end, back-end and database schema - appear in the IDE.
Data from the recent Sylogist Q3 2025 earnings call shows SaaS subscription revenue grew by 12 per cent year-over-year, yet the same report flags rising customer churn as developers migrate to in-house AI solutions (Sylogist). That tension is a clear indicator that the market is moving beyond the traditional SaaS model.
From a security standpoint, Solutions Review’s 2026 cybersecurity predictions warn that auto-generated code can introduce novel attack surfaces if not properly vetted. Irish firms are responding by embedding AI-security scanners into their pipelines, ensuring that the speed gains don’t come at the expense of safety.
For teams that have been shackled by a maze of SaaS integrations - think CRM, analytics, and payment gateways all wired together - the appeal of a single AI engine that can generate custom connectors is undeniable. Fair play to those who’ve built their business on a patchwork of plugins; the new approach simply offers a cleaner, cheaper alternative.
SaaS Plugins vs AI-Generated Code: A Side-by-Side Look
| Aspect | SaaS Plugin Stack | AI-Generated Code |
|---|---|---|
| Initial Setup Time | 2-4 weeks (config + integration) | 1-3 days (prompt + generate) |
| Ongoing Licence Cost | €12 000 / yr (multiple services) | €2 000 / yr (AI platform subscription) |
| Customisation Flexibility | Limited to plugin APIs | Full control over codebase |
| Security Audits | Vendor-provided reports | In-house automated scans |
| Scalability | Vendor-dependent limits | Elastic cloud deployment |
The numbers are illustrative, but they capture the shift in cost structure and speed. As the table shows, the AI route front-loads effort in prompting and validation, but it pays off quickly.
"The real power isn’t in cutting costs - it’s in freeing developers to solve problems that matter, not glue code," says Ciarán O’Leary, CTO of a Belfast fintech that recently switched to AI-first development (OpenTools).
That sentiment mirrors a broader industry observation: when the grunt work disappears, creativity returns. Teams can now allocate time to data analytics, user experience, and strategic innovation rather than spending hours on API mapping.
Irish Companies Leading the Change
Ireland has become a hotbed for AI-driven development, thanks to a supportive ecosystem of incubators, tax incentives and a talent pool steeped in both software engineering and data science. Companies like Quorum, which reported a modest 1 per cent revenue increase in Q3 2025, are quietly investing in AI code generators to offset the plateau in SaaS revenue (Quorum). Their CFO, Niamh, told me that the firm’s development budget has been cut by 15 per cent after adopting an AI platform that handles routine feature requests.
Another standout is a Cork-based health tech start-up that used Legato’s vibe builder to launch a patient-portal in just six weeks, a timeline that would have taken six months using conventional SaaS components. The founder, Dr. Eoin McDermott, noted, "We saved €200 000 in licence fees and got to market before any competitor could even think about it."
These success stories are echoed in the Netguru 2025 list of best AI SaaS solutions, which highlights several Irish providers that blend AI with traditional SaaS models, creating a hybrid that offers the best of both worlds. The list underscores that the market isn’t moving away from SaaS entirely; rather, it’s evolving into SaaS-plus-AI, where services are enhanced by on-the-fly code generation.
Fair play to the early adopters; their willingness to experiment is paving a path for the rest of the sector. As more firms see the tangible ROI - reduced licences, faster time-to-market, and a slimmer dev headcount - the 30 per cent myth will continue to dissolve into a richer narrative of AI-enabled development.
What This Means for the Future of Software Development
For Irish developers, the skillset shift is clear. Prompt engineering, model fine-tuning and AI-security testing are becoming as essential as knowledge of React or .NET. Universities such as Trinity are already introducing modules on generative AI, and I’ve seen bootcamps that teach "AI-first development" as a stand-alone track.
From a business perspective, budgeting will change. Instead of counting licences, CFOs will allocate spend to AI platform subscriptions and validation tooling. The 30 per cent figure will likely be re-interpreted as the proportion of the development budget that can be automated, a metric that will be tracked alongside traditional ROI.
In short, the shift isn’t a simple swap; it’s a transformation of the development lifecycle. The myth of a static 30 per cent move gives way to a dynamic, AI-driven ecosystem where SaaS plugins are just one option among many. As we move forward, the Irish tech scene is well-positioned to lead, blending home-grown talent with cutting-edge AI to deliver faster, cheaper and more secure software.
Frequently Asked Questions
Q: What exactly does the "30% SaaS vs Software shift" refer to?
A: It refers to the proportion of development effort that can now be automated with AI code generators, not a literal migration of 30 per cent of apps to SaaS. The figure highlights how much coding work can be replaced, freeing teams to focus on higher-value tasks.
Q: How much time can AI auto-code tools actually save?
A: In trials, teams have reported up to a 90 per cent reduction in coding time for routine features. Real-world case studies in Dublin and Cork show prototype delivery dropping from months to weeks, or even days.
Q: Are SaaS plugins becoming obsolete?
A: Not obsolete, but their role is shifting. Many organisations now use SaaS for niche services (payments, identity) while the core business logic is built with AI-generated code, creating a hybrid stack that reduces licence spend.
Q: What skills should Irish developers focus on next?
A: Prompt engineering, AI model fine-tuning, and AI-focused security testing are becoming core competencies. Traditional coding remains valuable, but the ability to guide and review AI-generated output is now a top priority.
Q: How will regulation affect AI-generated code?
A: The EU AI Act will require documentation and human oversight of AI-produced code. This means firms must keep audit trails, run security scans and retain accountability, which actually benefits those moving away from opaque SaaS vendors.
