Irish businesses looking to cut software spending by replacing established enterprise platforms with custom-built AI agents may be exposing boards to far greater legal and operational risk, according to Dublin consultancy Nucleo. The warning comes as companies continue to search for clearer returns from AI spending and reconsider large software licence bills.
The backdrop is a volatile technology market and a growing debate over what some in the sector have labelled a “SaaSpocalypse”. Nucleo said more than $285bn was wiped from global SaaS market capitalisations in a single week in February, after fears that AI agents could undermine traditional licence-based software models triggered a broader sell-off. For boards under pressure to free up budget for AI projects, cutting SaaS spend may appear to offer an immediate answer.
In Ireland, that pressure is colliding with uneven AI outcomes. According to the PwC 2026 Irish CEO Survey cited in the release, 51% of Irish leaders see the pace of technological change as their most pressing concern, while only 23% have so far achieved cost reductions from AI. That gap between expectation and realised value is encouraging some boards to look for savings by cancelling software subscriptions and moving towards internal AI tools.
Nucleo argues that the economics can be misleading. While custom AI agents may look cheaper on paper than enterprise software contracts, many organisations do not have the data architecture, governance, and operational controls needed to deploy them safely. In that scenario, cost savings on licences are offset by a transfer of responsibility from established software providers to the company itself.
Bobby Brown, founder and CEO of Nucleo, said: “Cutting enterprise licences looks great on the balance sheet for 2026, but it is a false economy. Most organisations simply do not have the rigorous data governance required to build bespoke AI agents. If your underlying data is messy, building a custom bot to run your operations is just a very fast way to automate your errors.”
That shift matters because established enterprise platforms generally come with defined support, vendor accountability, and risk frameworks. Building AI-driven workflows internally removes much of that protection. Brown said: “When you buy a software licence, you have a vendor to hold accountable. If a bespoke agent hallucinates a financial forecast or breaches GDPR, the company owns that error 100%. Under the EU AI Act, non-compliance can trigger fines of up to €35 million or 7% of global turnover. There is no vendor to support and nowhere to hide.”
The warning also extends to director oversight. Under Irish company law, Brown said, directors have a fiduciary duty to exercise “care and skill”. Approving custom AI systems without understanding the condition of the underlying data or the control environment around them could create personal exposure if governance failures later emerge.
Brown added: “The goal isn’t to avoid AI; it’s to avoid building a liability. Success starts with getting your data in order so you can deploy AI within the secure platforms your business already trusts.”
For Irish boards, the debate is therefore less about whether AI agents will become more common and more about where the accountability sits when things go wrong. In organisations with mature data governance, custom tools may still play a role. In less prepared environments, the cost of replacing SaaS could prove much harder to quantify than the licence bill it removes.
