AI & Project Portfolio Management: From Promise to Reality

For years, artificial intelligence applied to project management has been presented as an imminent revolution. Automation, risk anticipation, and augmented decision-making were among the numerous promises. In 2025, one thing is clear: AI is now widely adopted, but its real value depends heavily on the maturity of organizations and their project management practices.

In an increasingly complex Project Economy, where company performance depends on the ability to manage portfolios of projects, products, and transformations, the question is no longer “Should we use AI?” but rather “How can we integrate it in a way that is useful, measurable, and sustainable?”

Widespread Adoption, but Still Highly Uneven

The Second Global Research on AI in Project Management (2025), conducted among 870 professionals across 97 countries, shows a clear shift: 66% of professionals are already using AI tools in their project activities, compared to 41% in 2023. (Nieto-Rodriguez & Viana Vargas, The Second Global Research on AI in Project Management, 2025).

However, adoption remains largely partial. Only 12% of respondents report extensive and structured use. Most organizations rely on AI for isolated use cases, without a fundamental redesign of their project processes.

The challenge, therefore, is no longer access to AI, but the ability to absorb it into an existing project system.

I – From “AI Gadgets” to Useful AI: Where Value Is Really Created Today

While media narratives often focus on content generation or conversational assistants, the 2025 study reveals a far more pragmatic reality. The use cases perceived as most valuable are:

Planning and resource allocation (72%)
Predictive risk management (64%)
Financial and budget forecasting (41%)

In other words, AI is creating value at the very core of project execution, where timelines, costs, and decision credibility are at stake. This reinforces a key idea: AI is effective only when it relies on structured data, clear processes, and existing governance. Without these foundations, AI remains a tactical tool, sometimes impressive, but rarely transformative.

Measurable Business Impact Is Now a Reality

One of the major contributions of the 2025 study is the ability to quantify economic impact:

47% of organizations report cost reductions
39% report faster delivery timelines
26% report ROI or gains exceeding $250,000, with some surpassing $1 million

AI in project management is no longer an abstract promise; it is becoming a measurable performance lever when integrated coherently.

The Real Barrier Is No Longer Technology

A key finding: the nature of obstacles has fundamentally changed.

Obstacles	2023	Obstacles	2025
Employee resistance	66%	Employee resistance	62%
Lack of technical skills	62%	Lack of technical skills	48%
Ethical / transparency concerns	63%	Ethical risks / bias	54%
Budget	36%	Budget	33%

In 2023, barriers remained strongly intertwined with technical and organizational limitations, particularly a lack of technical skills and resistance to change. Two years later, while resistance and trust issues remain dominant, purely technical constraints have clearly diminished, indicating that technology is progressing faster than organizations’ ability to integrate it effectively.

This confirms a frequently underestimated reality: AI does not transform organizations on its own. It amplifies what already exists, whether it is best practices or dysfunctional uses.

II – The Planisware Perspective: AI Serving the Project Economy, Not a Trend

At Planisware, this reality is not new. As Pierre Demonsant, Founder and Chairman of Planisware, explained, AI has been part of the company’s DNA since its origins, long before the current wave of generative AI.

“We started working on AI in the 1990s, on expert systems and early machine learning. From the creation of Planisware, we knew AI would eventually become a key lever for managing projects.”

As a result, at Planisware, AI is not designed as a standalone tool or a generic assistant. It is embedded directly in project data, governance processes, and the real operational constraints of portfolio management. Whether predictive or language-based, AI capabilities leverage project histories, planning structures, governance rules, and financial data to support anticipation, arbitration, and decision-making, rather than generating recommendations disconnected from reality.

AI That Understands Projects, Not Just Language

Large Language Models have unlocked new possibilities but also revealed clear limits. As Pierre Demonsant notes:

“LLMs are extremely powerful for language, but portfolio management is also about durations, costs, dependencies, and resources. The real challenge is linking textual meaning with structured data.”

In this context, Planisware has developed capabilities that enable:

Semantic analysis of exchanges (emails, discussions, task descriptions)
Detection of weak signals (issues, risks, tensions)
Dynamic structuring of unstructured data (auto-tagging, activity categorization)
Feeding reliable predictive models based on data quality

A Core Conviction: No High-Performing AI Without Project Maturity

Planisware’s message is clear: AI does not replace project maturity; it depends on it. Predictive models, intelligent assistants, and advanced analytics create value only if:

Data quality is sufficient.
Processes are clearly defined.
Project governance is under control.

That is why Planisware invests heavily in:

Data quality: AI is never better than the project data it is fed. In the project management field, data is often incomplete, inconsistent, or heterogeneous. Building predictive or decision-support systems on poor data simply industrializes error. AI should not assume data quality; it must first challenge it.
Explainability: Many AI tools produce answers without explaining which data was used, which criteria were weighted, or how outcomes would change in different contexts. In project environments, a non-explainable decision is an unusable decision. Risk alerts must be linked to their triggers; prioritization suggestions must clarify trade-offs between value, risk, and capacity.
AI as a copilot, never an autonomous decision-maker: In Project Portfolio Management, decisions are multi-factorial. They involve non-modelable elements (internal politics, strategy, market context) and always carry human accountability.

III – Why Planisware Is Launching Its Own AI & PPM Barometer

Global studies show that AI adoption is accelerating. But they do not answer a critical question for organizations: Where do you really stand, within your projects and portfolios, compared to others in your market?

This is the transition Planisware aims to clarify through its own study: from experimentation to maturity. The objective is straightforward:

Measure the real level of AI maturity in project and product management.
Identify what still belongs to experimentation versus what already creates value.
Enable organizations to benchmark themselves objectively against peers, by industry, and project maturity level.

In just a few minutes, your participation helps build a reliable market benchmark and will provide you, once results are consolidated, with actionable insight to better guide your AI decisions.