Businesses Risk AI Setbacks As Legacy Systems Struggle To Keep Pace, New Study Finds
By John O’Brien, CEO NashTech
Businesses risk seeing their AI ambitions stall as weaknesses in existing technology systems are exposed, according to new global research from NashTech.
The findings are published in Differentiating through custom software: The NashTech 2026 report on software development in the AI age, which explores how custom software and commercial off-the-shelf solutions shape organisations’ ability to scale AI effectively.
The study shows that 96% of organisations say AI is accelerating changes in technology strategy and 85% describe it as an immediate priority for custom software development (CSD). Yet as AI moves from experimentation to deployment, integration is emerging as a critical barrier, and foundational weaknesses are beginning to surface.
While 44% of respondents say they invest in custom software primarily to improve integration, 40% identify it as their biggest challenge, and 47% say legacy integration could impact compliance. This suggests that architectural limitations are becoming enterprise risk issues.
The results are drawn from a global survey of 1,000 technology decision-makers conducted by independent research firm Vanson Bourne across EMEA, North America and APAC.
“The AI conversation has been dominated by models and use cases. But in most enterprises, the real constraint is not intelligence. It is integration,” said John O’Brien, CEO at NashTech. “AI does not fail because of algorithms. It fails because the systems beneath it were never built for intelligence at scale. What we are seeing is integration debt quietly turning into AI debt.”
AI ambition versus operational readiness
The research suggests most organisations recognise AI’s transformative potential. Three-quarters expect AI to have a significant positive impact on custom software, from accelerating development cycles and automating testing to improving code quality and enabling adaptive, data-driven systems. Meanwhile, 85% say they have begun or will begin adopting AI within the next 12 months, with most also moving to formalise AI governance.
However, adoption exposes structural gaps. Data privacy across systems is cited as a top risk by 49%, while 48% express concerns about third-party handling of sensitive data.
Governance practices vary significantly. Although organisations report using a mix of controls and mitigation strategies, only 47% say they use board-level reporting as part of their risk management approach.
The findings suggest many enterprises are still laying the necessary groundwork in data, integration and governance before scaling advanced generative and agentic AI capabilities.
The perception gap at the top
The report also identifies a growing disconnect between executive confidence and delivery reality.
While 63% of senior leaders say their CSD projects exceeded expectations, only 39% of mid-level managers agree. Those closer to implementation are more likely to report operational friction, with 36% citing scope creep and 46% pointing to integration challenges.
This perception gap risks masking delivery strain until projects scale, particularly when optimism at the top is not matched by operational readiness on the ground.
Quality remains the priority
Despite pressure to move faster, engineering discipline continues to outweigh speed. 46% say they aim to balance speed and quality, but when forced to prioritise, quality wins, particularly at senior levels.
Rethinking partnerships for long-term value
As AI increases architectural and governance complexity, expectations of technology partners are shifting.
While 47% of organisations describe their partners as trusted delivery providers, only 32% consider them truly strategic. Yet appetite for deeper collaboration is clear: 97% say they would be willing to invest in partners that consistently deliver long-term value.
The full report is being unveiled today at NashTech Connect in London and is available to download at: The State of Custom Software Development 2026 report
About NashTech
NashTech is a global provider of custom software development, technology advisory services, training services for AI models and application managed services. It develops custom-built differentiated solutions for enterprise customers and works with independent software vendors to design and build innovative new software products.
The company operates across 11 offices in nine countries, supported by five delivery centres and a team of more than 2,000 engineers. NashTech works with organisations across all sectors and has particularly strong industry specialisation in Insurance, Financial Services, Retail and Consumer Products, Logistics and Higher Education, with clients including Evri and Hays Travel.
About the study
NashTech commissioned independent research company Vanson Bourne to conduct a survey of 1,000 technology decision-makers across six key territories in 2025, spanning organisations with 500 to 9,999 employees across EMEA, North America and APAC. The research examined strategy and prioritisation, delivery discipline, integration and compliance, partner value, AI adoption and (for ISVs) product-centric scaling factors.