Faisal Saeed is the Founder & CEO at Promptev Inc. At the forefront of AI’s next leap.
While enterprises pour record investments into artificial intelligence, a critical infrastructure gap is quietly sabotaging billions in potential value. The numbers tell a sobering story. Worldwide generative AI spending is expected to hit $644 billion in 2025, up 76.4% from last year, according to Gartner. Yet MIT’s NANDA initiative found that roughly 95% of generative AI pilots at companies are failing to deliver their promised value.
What’s going wrong? While boardrooms celebrate flashy AI demonstrations and chase the latest large language models, they’re missing the operational foundation that separates successful deployments from expensive experiments. The gap isn’t in the technology itself; it’s in the infrastructure.
The Hidden Operational Crisis
McKinsey research points to five major challenges facing AI adoption: aligning leadership, addressing cost uncertainty, workforce planning, managing supply chain dependencies and meeting demands for explainability. But there’s a more fundamental issue underneath: how companies actually manage enterprise context-driven prompt engineering and the workflows powering their AI systems.
Right now, the instructions guiding AI and agentic models (commonly known as LLM prompts and used for everything from customer service to financial analysis) exist without the operational discipline applied to traditional software. Most organizations lack prompt version control tools to track changes, systematic testing before deployment and clear governance over who can modify critical AI behaviors.
Industry experts emphasize the need for rollback mechanisms and audit trails and safety nets to trace and fix issues quickly, yet most companies don’t have these basic capabilities in place. McKinsey found that while almost all companies invest in AI, just 1% believe they’re at maturity. That 99% gap represents billions in unrealized value.
The Compliance Pressure Cooker
The pressure is intensifying fast. The EU AI Act’s first major deadline hit on February 2, 2025, with comprehensive requirements taking effect on August 2. Noncompliance isn’t cheap, with penalties reaching up to 35 million euros or 7% of worldwide annual turnover, whichever is higher.
For global enterprises, this means AI systems deployed anywhere must meet rigorous standards. Organizations need answers: Who approved this decision? What data did it access? How was it tested? Without proper prompt life cycle management platforms, these questions become impossible to answer.
Why 2025 Changes Everything
PwC reports that company leaders can no longer address AI governance inconsistently. As AI becomes intrinsic to operations, companies need systematic governance through AI workflow automation. Gartner identifies AI agents as one of the fastest advancing technologies in 2025, signaling the shift from simple chatbots to autonomous systems that make decisions and take actions.
The scale is staggering. Organizations are orchestrating hundreds or thousands of prompt engineering workflows across departments, each potentially accessing sensitive data and impacting customer experiences. Companies investing in AI adoption are expected to contribute $19.9 trillion to global GDP by 2030, but capturing that value means treating AI deployment with the operational discipline that revolutionized software development.
Generative AI’s DevOps Moment
Think back 20 years. Companies shipped software without version control or continuous integration. DevOps practices transformed that chaos into systematic delivery. Cloud infrastructure turned computing from a capital expense into an on-demand utility.
Generative AI is hitting its own operational awakening. The solution mirrors what worked for software: prompt version control, testing environments that catch issues before production, governance layers ensuring AI compliance through prompt governance and observability tools providing visibility into behavior.
Scalable, AI-native infrastructure now enables enterprises to deploy in days what once took months to build and maintain. The real work ahead isn’t about chasing the next model breakthrough; it’s about integrating robust, proven operational frameworks that support governance, reliability and scale. The organizations that master this foundation will be the ones to translate experimentation into measurable enterprise success.
The Real ROI Story
Every dollar invested in generative AI delivers an average 3.7 times ROI, but only when systems actually work reliably. Many organizations have already decided to integrate specialized AI platforms or prompt management infrastructure to scale their initiatives. For them, success now depends on how they execute.
Common challenges often arise during integration, unclear ownership, fragmented workflows or limited testing processes. Companies need certain foundations in place, like clear governance frameworks, data quality standards and alignment between technical and business teams. Enterprises that address these early by establishing version control, testing sandboxes and defined approval hierarchies tend to realize faster time-to-value and fewer downstream risks.
I suggest setting measurable success criteria, embedding auditability from the start and fostering collaboration between AI developers, compliance teams and business units. This balanced approach ensures that the platform investment delivers sustained value rather than one-off wins.
With global private investment in generative AI having reached $33.9 billion in 2024, capital is increasingly gravitating toward foundational infrastructure—the essential enablers of the AI ecosystem.
The Closing Window
With worldwide generative AI spending approaching $644 billion, every percentage point can represent billions in value. The future won’t be decided by who has the most sophisticated models or the largest training budgets. I believe it will be decided by who can turn experimental technology into reliable, governed and scalable systems delivering measurable value day after day.
That transformation is happening right now. The only question is whether you’ll be leading it or watching from behind.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
