Legacy Systems Halt 84% of Enterprise AI Projects as Infrastructure Gaps Threaten Returns
Eighty-four percent of technology leaders at large global companies report that legacy system limitations have forced their organizations to cancel artificial intelligence pilots or projects, according to a recent study by GFT Technologies. The research, which surveyed 945 chief information officers and chief technology officers across 19 countries at companies with annual revenues exceeding US$500 million, highlights a growing disconnect between ambitious AI adoption strategies and the outdated technological foundations required to support them.
Infrastructure Gaps Outpace Enterprise Software Budgets
Organizations have rapidly deployed generative AI models, automation, and copilots onto information architectures built decades before modern machine learning existed. A typical enterprise often fragments its operations across siloed environments, storing customer data in one platform, financial records in another, and legacy software in older layers. Each system operates with distinct access rules, security protocols, and data structures. Marco Santos, global CEO of GFT Technologies, noted that legacy infrastructure has become a real constraint for innovation, security, and scalability as companies rush to implement new tools.
This structural friction directly impacts financial returns. Eighty-nine percent of surveyed technology leaders express concern that global AI investments are accelerating faster than the actual business value generated. Maintaining incompatible platforms, executing manual data integrations, and moving information between disparate systems before a model can process it drives up the cost and complexity of every deployment.
Security Risks and Modernization Pressures Facing Global Businesses
Failing to modernize underlying architectures before scaling enterprise AI introduces severe vulnerabilities. The GFT study reveals that 93% of surveyed technology leaders believe running artificial intelligence on unprepared legacy systems will eventually trigger an enterprise-scale security crisis. Consequently, technology modernization is shifting from a standard efficiency upgrade into an essential risk-reduction mandate.
Rather than executing a total and costly replacement of every legacy asset, many large organizations are opting for targeted updates. Modernization strategies frequently involve migrating specific applications to cloud environments, updating architectures, or converting older systems into real-time information exchange platforms. This approach bridges the operational gap between legacy software maintenance and modern AI requirements.
Latin American Enterprises Confront Complex Integration Hurdles
The infrastructure challenge carries particular weight in Latin America. Banks, insurance providers, and telecommunications firms throughout the region have spent years advancing digital transformation initiatives, accumulating multiple generations of technology along the way. Introducing artificial intelligence layers on top of these complex architectures deepens operational challenges.
Technology executives must now evaluate whether their existing systems can reliably supply AI models with the secure data connections and operational controls necessary for production deployment. According to GFT’s findings, companies that successfully turn AI investments into measurable outcomes are those approaching the technology as part of a comprehensive transformation encompassing data governance, security, business strategy, and infrastructure overhauls.
Frequently Asked Questions About Enterprise AI Infrastructure
What specific company size did the GFT Technologies study analyze?
The study surveyed 945 chief information officers and chief technology officers at organizations with annual revenues exceeding US$500 million across 19 countries.
Why do older legacy systems interfere with modern artificial intelligence tools?
Legacy systems often store customer data, financial transactions, and inventories across isolated platforms with incompatible data structures, requiring manual integrations and complex data transfers before an AI model can utilize the information.
How are large enterprises modernizing their technology without replacing all legacy systems?
Instead of executing complete system replacements, companies are migrating specific applications to the cloud, updating architectures, and turning legacy systems into platforms capable of real-time information exchange.