Scale at Speed™
English (Global)
French (Canada)
- The disparity between AI adoption and decision-making quality is rooted in pillar misalignment rather than the tools used.
- Organizations that progress on several UPM pillars simultaneously tend to achieve better results than those that focus on just one.
- The majority of portfolios respond 'no' to at least two out of three diagnostic inquiries, revealing that Level 2 foundations are present beneath Level 4 tools.
The CIO Scoreboard Has Changed
A chief information officer at a FTSE-100 company asks her leadership team a simple question: what's our AI strategy for portfolio management?
The team has the answer. Over the past two quarters, three enterprise-level AI capabilities have gone live within the project portfolio management (PPM) framework. Pilot initiatives are already underway.
The CIO listens. Then she points out that every answer she just heard pertains to tools. That wasn't her question.
Similar conversations are playing out across industries. We've sat in versions of this meeting across pharma, banking, and telecom over the past nine months, and the pattern repeats. The number of AI features on PPM platforms continues to grow. But it is not translating to better portfolio decisions.
Industry research reinforces this trend. McKinsey reports that 88% of organizations now use AI in at least one business function, yet only one-third have initiated enterprise-wide scaling.1 BCG's September 2025 study of 1,250 companies paints a similar but more concerning picture: 60% of firms aren’t generating any significant AI value, while merely 5% have achieved value at scale.2 The challenge is ensuring that AI changes how decisions are made.
AI adoption isn't the gap anymore. It’s the disconnect between adoption and the decisions a portfolio committee can act on.
Why More AI Features Don't Produce Better Decisions.
When a portfolio team says it has switched on an AI assistant, it has answered a question about tools. It has said nothing other than that.
It hasn't said whether the criteria behind investment decisions have moved beyond weighted-scoring rubrics, whether the governance can act on AI-led insights, or whether policies can justify AI-influenced investment decisions to a regulator.
Six factors determine whether a portfolio function decides well: process, tools, budgeting, governance, decision criteria, and policies. Over the past two years, we worked with portfolio leaders on this problem and built a framework around those six, called Unified Portfolio Management. Beneath the six lie 13 modules and 60 capability items, each assessed on a five-level scale:
- Level 1: No capability
- Level 2: Manual
- Level 3: Standardized
- Level 4: AI-augmented
- Level 5: Autonomous
The same pattern keeps repeating. Teams buy Level 4 tools and run them on decision criteria, governance, and policies that are still at Level 2.
That mismatch is the growing concern, and more technology doesn't solve it. Without redesigning what sits underneath, AI increases activity without reducing effort. It expands monitoring without strengthening control. Analysis expedites without improving responsiveness.
A team can have an advanced tool that forecasts, simulates, flags risk, and still decide the old way, because the criteria feeding it are manual and the governance receiving its output was built for status reporting. The Level 4 tool ultimately produces Level 2 decisions.
The AI features keep coming up in PPM platforms. But the quality of portfolio decisions doesn’t move.
Same technology. Different decisions.
The question facing CIO leadership teams now: what decisions are possible today that weren't a year ago, and what evidence supports them?
For instance, let’s take two clients who ran pilots in the same quarter with comparable scope, tools, and spend. Eighteen months later, one portfolio committee makes investment calls faster, on evidence its peers can't produce, and responds to signals its competitors do not see. In contrast, the other committee still convenes monthly meetings to review the same status report. The AI quietly operates in the background with little to no impact on portfolio decisions.
Different outcomes arose because one organization integrated its UPM tools and decision criteria pillars, whereas the other advanced only its tools pillar. The former evaluated the AI's results, weighed them against priorities, and documented the reasoning behind every call.
This is what Level 5 actually looks like. Supported with better evidence, the committee decides faster. Every decision has a record behind it. And the governance forum has been rebuilt to work with that evidence, instead of running through a monthly status pack.
Every funding decision has a record behind it, and a committee that could defend it a year later.
The CIO's job is simpler now: find the pillar that's holding them back, and move it.
Three Questions for Monday Morning
Finding that tool takes three questions, each testing a different foundation/corresponding to a UPM pillar.
If the answer to any of them is no, the AI conversation in your portfolio function is set to Level 2. That has to be rectified before the next project.
- Process and Tools: Are your portfolio decisions running on data that is current as of the last steering committee, deduplicated, and traceable to a system of record? If the answer is no, your tools are holding your process back.
- Decision Criteria and Governance: When the steering committee makes a portfolio decision, is the rationale documented with an explicit weighting the committee could defend to an auditor a year later? If not, your decision criteria are a scoring rubric rather than a decision-making instrument.
- Policies: If a regulator asked tomorrow how AI shaped your last three portfolio decisions, could you provide the evidence trail within a single working day? If the answer is no, it implies that the policies pillar is at Level 1 or 2, meaning that the AI utilized in your portfolio is operating outside your audit boundary.
In total, there are three questions addressing six pillars. Allocate five minutes for each portfolio function.
In our assessments of various portfolios, at least two of the three questions are likely to receive a negative response. The instinct is often to invest in the most prominent tools, as that is where vendor hype is. But the real advantage lies in the pillars of either decision criteria, governance, or policies. The pillar impeding your progress is rarely the one vendors are emailing you about.
Back to the Boardroom
The CIO in our opening scene wasn’t being obstructive. She was conveying to her committee, in fewer words than this article, that the AI conversation in her portfolio function has overlooked a crucial aspect. Tools are a question about the Tools pillar. Decisions encompass the entire framework.
In the coming 12 months, the gap between portfolios that possess the framework and those that only have the features will increase. Both will spend on AI. However, only one will be able to inform their CEO what that investment has delivered.
This article marks the beginning of a 13-part series focused on AI and Unified Portfolio Management. The subsequent article will explore why the majority of AI projects falter before reaching any conclusions, as well as what the failure rates indicate about which UPM pillar should be prioritized for improvement.
TAGS:
Artificial Intelligence
Strategy and Consulting
Frequently Asked Questions
Our FAQ section is designed to guide you through the most common topics and concerns.
AI has significantly improved the speed and depth of analysis, but many organizations continue to make portfolio decisions using governance models, review cycles, and decision criteria designed for a pre-AI environment. As a result, AI often generates more insights without changing how investment decisions are evaluated, approved, or acted upon. The gap is no longer AI capability—it's the organization's ability to integrate those capabilities into its decision-making processes.
The most common challenge is treating AI as a technology initiative rather than an operating model change. Organizations often invest in AI-enabled portfolio tools without updating governance, decision criteria, budgeting practices, or policy frameworks. When these capabilities evolve independently, AI enhances visibility but has limited impact on the quality or speed of portfolio decisions.
Unified Portfolio Management (UPM) is a diagnostic framework developed by Tech Mahindra to help organizations assess the capabilities that influence portfolio decision-making. Rather than focusing solely on AI tools, it examines how processes, governance, budgeting, decision criteria, policies, and technology work together to support faster, evidence-based, and more accountable investment decisions.
A useful starting point is to evaluate whether portfolio decisions are based on trusted, current data, whether decision rationale is transparent and auditable, and whether governance processes are equipped to act on AI-generated insights. These questions often reveal whether the organization's decision-making capabilities have evolved alongside its AI investments.
Rather than adding more AI capabilities, organizations should first identify the constraints limiting decision quality. In many cases, the bottleneck lies in governance, decision criteria, policy, or process rather than technology. Addressing these foundational capabilities enables AI investments to deliver greater business value and improve portfolio outcomes.
