By Richard 
Enterprise AI has quickly transformed from a promising technology to a competitive necessity. However, many CIOs and IT leaders are facing increasing pressure from their enterprise software vendors to migrate to cloud platforms to access new AI capabilities – even when their current on-premises systems are stable, customized and meeting business needs. Cloud migrations are expensive, complex and disruptive—so why is it often presented as the only path to realizing the benefits of AI?
While cloud platforms offer advantages, mandated cloud migrations solely to enable AI functionality do not always meet ROI assessments of such undertaking and can create significant challenges that IT leaders must carefully evaluate.
Challenge #1: Innovation at the Vendor’s Whim
Many organizations have spent years developing sophisticated, customized on-premises systems that run their core business processes efficiently. When vendors limit new AI capabilities to cloud-only offerings, they effectively constrain their customers’ ability to innovate at the speed their business demands.
The reality is that different AI providers excel in different areas—some might offer superior natural language processing, while others lead in predictive analytics or computer vision. By limiting AI implementation to a single vendor’s cloud ecosystem, organizations risk missing opportunities to adopt more advanced or specialized AI solutions that better fit their specific use cases.
And if we learned anything from SAP’s abrupt announcement that innovations will only be available for cloud customers, even for those who moved to S/4HANA on-prem thinking they were upgrading to be able to access the latest innovations, is that vendors can change the game anytime – and you may be the one having to go back to your board, hat in hand.
Challenge #2: Loss of Strategic Control
Cloud migration involves more than technical changes—it fundamentally shifts how organizations manage and pay for their software infrastructure. Moving from owned, perpetual licenses to subscription-based models can impact long-term costs and negotiating leverage.
Moving to a vendor’s cloud platform often means surrendering certain aspects of control over your IT infrastructure and data. Organizations may find themselves locked into specific feature sets, upgrade cycles, and pricing models, potentially limiting their ability to adapt quickly to changing business needs.
IT leaders should carefully evaluate the total cost of ownership for cloud-based AI initiatives, including hidden costs like data transfer fees, storage costs and potential premium pricing for AI-specific features.
Challenge #3: The Importance and Value of Historical Data
AI systems require extensive amounts of clean, historical data to deliver accurate insights and predictions. However, many organizations migrating to cloud platforms face a difficult choice: Leave behind years of valuable historical data or pay the hefty price to migrate and store it in the cloud.
Many companies end up moving only a few years of data, leaving behind decades of invaluable information and context, which can significantly impact the effectiveness of AI algorithms.
Challenge #4: Data Silos and Limited Scope
Modern enterprises maintain data across various platforms, including specialized departmental applications, IoT devices and external data sources. Enterprise AI implementations deliver the most value when they can analyze data from multiple sources across the organization—not just from a single system.
Cloud-only AI offerings from enterprise software vendors typically focus on data within their own ecosystem, creating potential blind spots in AI analysis by missing valuable insights from other enterprise systems and data sources.
A Flexible, Future-Ready Approach to AI
Rather than viewing cloud migration as a prerequisite for AI adoption, organizations can consider a more flexible, future-ready approach:
- Focus on Data Accessibility: Rather than moving all data to the cloud, implement data orchestration layers that make information accessible to AI systems regardless of where it resides. This approach preserves valuable historical data while enabling advanced analytics and AI capabilities.
- Adopt a Composable Strategy: Implement a “composable” approach that allows IT leaders to integrate best-of-breed AI solutions while maintaining core systems. This enables innovation around the edges of existing infrastructure.
- Prioritize Business Outcomes: Instead of following vendor-dictated roadmaps, develop AI strategies that align with business objectives. This might mean starting with smaller, more focused AI implementations that deliver immediate value rather than comprehensive platform migrations.
AI on Your Own Terms
While cloud platforms can offer valuable capabilities for AI implementation, they are by no means the only path forward. By carefully evaluating options and maintaining focus on business objectives, organizations can develop AI strategies that leverage their existing investments while positioning themselves for future innovation.
About the Author
Saulo Bomfim has over 30 years of experience leading high-performing global teams delivering products and services ranging from enterprise applications to emerging technologies and solutions. In his role as Vice President, Product and Service Strategy at Rimini Streethe is responsible for innovation and value creation for clients in the form of products and services that optimize, evolve, and transform their applications and technology ecosystems.
Sign up for the free insideAI News newsletter.
Check us out on YouTube!