Mohit Gupta, the CEO of Damco Solutions, is a visionary business leader with over 30 years of industry experience.gettyThere is a conversation happening in boardrooms and technology leadership meetings across every major industry. It follows a logic that feels airtight: The core system is IBM i, but IBM i is legacy, therefore the platform must be exited before the organization can meaningfully pursue AI.The reasoning is clean, but there is a possibility that this transition goes wrong, at a cost that far exceeds the problem it purports to solve.This is not a case for inertia. The modernization pressures bearing down on CIOs are real. Boards are demanding AI readiness on timelines that leave little room for deliberation. Capital budgets cannot absorb the full weight of a platform migration without crowding out other strategic priorities. The report program generator (RPG) talent pool is contracting, and the generation behind it is not replenishing it. These are legitimate constraints.Yet, through all of it, the platform runs well in most cases, processing the transactions the enterprise depends on every hour, every day.In some cases, rigorous examination may highlight complete migration as the genuine need of the hour. In many others, the instinct under sustained pressure is to treat the system itself as the problem.Before that instinct hardens into strategy, it deserves a deeper look.The Legacy Label Obscures More Than It RevealsWhen technology leaders describe IBM i as legacy, they are pointing to something real: an interface layer that has not kept pace with modern expectations, limited native exposure to AI and data tooling and business logic that lives inside the system rather than being accessible to it.What the legacy label does not accurately describe is the platform's operational character. IBM i is one of the most stable, high-throughput transaction processing environments ever built, offering renowned backward compatibility.​​The need to modernize arises from two pressures: the growing inaccessibility of embedded business logic to modern tooling, and the narrowing pool of professionals who understand it. These are integration and talent challenges, not platform failure.Organizations underestimating legacy code complexity are likely to run a high risk of experiencing program delays, with undocumented business logic acting as the single most common source of unplanned rework. Organizations that proceed without first surfacing and mapping it are not managing a modernization program. They are managing a discovery process at enterprise cost and under enterprise time pressure.​The Question Around Migration CostThe full migration path appears attractive until it is modeled with precision. Quite often, large-scale platform modernization programs exceed their original budgets, with only 30% of them fully meeting expectations for timeline, budget and scope, according to BCG research.A full migration may also consume the entire capacity of the IT organization during execution, leaving nothing for the innovation work that was the original rationale for the move.​One cost-efficient alternative to a full migration is to build an intelligent layer around IBM i. In this scenario, the core continues to perform what it does best while an AI and integration layer, APIs, machine learning models and workflow automation can deliver outcomes measurable from the point of implementation.​​ ​​​Why The Skills Gap Is Not Always A Cause For Migration ​The skills shortage in the IBM i ecosystem is measurable and accelerating, with 69% of the 315 IBM i users surveyed by Fortra citing being able to find technologists with the necessary skills for the platform as a top concern. ​However, accelerating migration can compound the problem. A rip-and-replace program requires a capacity-constrained organization to simultaneously manage a complex migration, maintain operations and build competency on an entirely new platform.What the skills gap calls for is reducing dependency on tribal knowledge, beginning with making the existing system legible. The business logic in RPG and control language (CL) code is often not understood by anyone currently in the organization. Before any meaningful modernization can proceed, that logic needs to be surfaced, documented and made accessible.What AI Readiness Requires​Three conditions must be true for IBM i to function as an AI-ready core: the data must be accessible, the business logic must be documented, and the team must be ready to operate a hybrid architecture. None of this requires replacing IBM i.​​The window for getting this right is narrowing. Competitors are not waiting for platform decisions before deploying AI.Full platform exit remains the right answer for some organizations, under the right conditions, with the right runway, subject to a thorough assessment. But first, organizations should consider if they can make the existing system an active participant in an AI-driven operation rather than a barrier to one. That means a rigorous audit of embedded business logic, evaluating AI discovery platforms and API integration layers and treating the extension layer as the primary delivery vehicle for AI capability.​​​​​Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?