De lo predictivo a lo ágil: hacia un marco híbrido en proyectos de I+D+i

Abstract

This article provides a conceptual comparison of predictive, agile and hybrid approaches for managing research, development and innovation (R & D & I) projects, with explicit attention to differences between low-TRL research projects and high-TRL innovation and transfer projects. Building on the R&D&I life cycle proposed by Carranza Ruiz de Loizaga, the paper develops a phase-dependent hybrid framework that integrates practices from PMBOK, P3.express, Scrum and Lean Startup to address a central tension in R & D & I management: enabling rapid learning under high uncertainty while ensuring governance, traceability and accountability to sponsors and regulators.The contribution is threefold. First, the article clarifies the differential value of predictive and agile approaches according to project phase and key attributes, including technological maturity (TRL), regulatory criticality, stakeholder complexity and time pressure. Predictive standards are shown to be particularly effective in phases where formalisation, quality assurance and configuration control are critical, notably during initiation, planning and transfer. By contrast, agile approaches provide greater value during exploration, prototyping and validation, where short feedback cycles and incremental experimentation reduce technical and market uncertainty.Second, the paper operationalises hybridisation through an explicit decision architecture that combines strategic decision gates with tactical iterative cycles. Decision gates define transparent entry and exit criteria related to technical maturity, desirability, economic viability and residual risk, while iterative cycles (e.g., sprints) are used to generate the evidence required to support each decision. This “gates plus iterations” logic helps prevent the confusion of activity with progress and aligns resource commitments with validated learning.Third, the article proposes a dual-metric system that complements traditional performance indicators (cost, schedule and quality) with learning-oriented metrics such as validated hypotheses, evidence quality and residual risk. This dual logic provides a shared evaluation framework for project teams, sponsors and industrial partners. Lightweight documentation practices, inspired by P3.express, are incorporated to preserve decision traceability without constraining experimentation.The paper concludes by discussing implications for R&D&I project governance and portfolio management. The proposed framework is conceptual and requires empirical validation across sectors, TRL levels and types of university–industry collaboration.