AI-Powered Project Portfolio Intelligence

Know what to build before you build it.

Organisations rarely suffer from a shortage of ideas.

They suffer from a shortage of capital, people, time, technology, expertise and attention to pursue all those ideas simultaneously.

A company may have 20 potential projects, 10 promising technologies, 5 strategic priorities and a limited budget. The difficult question is not simply:

Which project is good?

It is:

Which combination of projects should we pursue, given our objectives, resources, risks and constraints?

ProjectML is being developed to answer that question.

ProjectML is an AI-powered Project Portfolio Intelligence platform designed to help organisations evaluate potential projects, predict outcomes, simulate scenarios, optimise portfolios and make better project investment decisions.

The Problem

Good projects compete for scarce resources.

Most organisations evaluate projects individually.

A project may have:

strong revenue potential
attractive technology
strategic importance
promising market demand
an experienced team
acceptable risk

Yet that does not necessarily mean it should be selected.

The organisation may already have another project consuming the same engineers, capital or infrastructure.

Two individually attractive projects may be a poor combination.

A technically excellent project may be strategically premature.

A high-return project may expose the organisation to unacceptable concentration risk.

A project with lower expected returns may actually be essential because it creates capabilities required by several future projects.

This is the difference between project evaluation and portfolio intelligence.

Traditional project selection often relies on spreadsheets, scoring models, financial metrics, expert judgement and manually weighted criteria.

These approaches remain valuable, but increasingly complex portfolios require decision frameworks capable of handling:

multiple criteria + uncertainty + dependencies + resource constraints + competing objectives.

Research in project portfolio selection has extensively explored multi-criteria decision analysis and its combination with optimisation methods for resource-constrained portfolio decisions.

ProjectML is being designed around this problem.

The ProjectML Insight

A good project isn’t necessarily the right project.

And:

The best project isn’t necessarily the best portfolio.

ProjectML therefore shifts the question from:

“Should we do this project?”

to:

“What should we do with the projects we have?”

The platform is designed to evaluate individual projects and understand how projects interact within a portfolio.

What ProjectML Does

ProjectML follows a five-stage decision intelligence workflow:

01 — Evaluate

Assess projects across multiple dimensions.

Examples include:

Strategic alignment
Market potential
Financial attractiveness
Technology readiness
Resource requirements
Organisational capability
Execution complexity
Risk
Sustainability
Time-to-value

02 — Predict

Use historical data, project characteristics and machine-learning models to estimate potential outcomes.

Depending on available data, the system could eventually estimate:

probability of project success
expected schedule performance
cost overrun probability
expected return
delivery risk
resource pressure
technology risk
likelihood of strategic alignment

The objective is not to replace managerial judgement.

It is to provide better evidence for managerial judgement.

03 — Simulate

What happens if the assumptions change?

ProjectML can be designed to evaluate alternative scenarios such as:

10% budget reduction
loss of key resources
delayed market entry
increased project cost
reduced expected revenue
technology uncertainty
changed strategic priorities
different risk tolerance

This creates a what-if environment for project portfolios.

Instead of asking:

“What happens if we choose Project A?”

decision makers can ask:

“What happens if we choose A + C + F instead of B + D + E?”

04 — Prioritise

ProjectML converts multiple quantitative and qualitative factors into structured project and portfolio priorities.

The underlying research framework can incorporate multi-criteria decision-making (MCDM/MCDA) approaches, including methods such as AHP and TOPSIS, alongside predictive models and optimisation techniques.

This is particularly relevant because project portfolio decisions inherently involve competing criteria and stakeholder preferences.

05 — Optimise

The ultimate objective is not simply to rank projects.

It is to identify a portfolio that produces the strongest combination of value while respecting constraints.

For example:

Maximise

strategic value
financial value
market opportunity
innovation value

while considering:

Constraints

budget
people
skills
infrastructure
time
risk capacity
technology readiness
project dependencies

This portfolio-level perspective is fundamental to project portfolio selection research.

From Project Score to Portfolio Intelligence

Traditional approach:

Project A → Score 82

Project B → Score 79

Project C → Score 76

Therefore:

Select A, B and C.

ProjectML aims to ask a more sophisticated question:

Can A, B and C collectively be executed with the available resources?

Perhaps:

A + B + C

creates resource conflicts.

But:

A + C + F

may create greater overall portfolio value.

This is where ProjectML moves beyond project scoring.

ProjectML Decision Architecture

The conceptual architecture is:

INPUT

Project Data

Project objectives
Cost
Duration
Resources
Expected revenue
Technology
Market
Risk
Strategic alignment
Dependencies

↓

INTELLIGENCE LAYER

ProjectML Engine

Data processing
Feature engineering
MCDM
Machine learning
Predictive analytics
Scenario modelling
Monte Carlo simulation
Portfolio optimisation
Sensitivity analysis

↓

DECISION LAYER

Portfolio Intelligence

Project score
Success probability
Risk profile
Resource utilisation
Portfolio combinations
Scenario comparison
Portfolio recommendation

↓

OUTPUT

Decision Support

Select
Defer
Reconsider
Reject
Monitor

with an explanation of why.

Human + Machine

ProjectML is not intended to become a black-box system that tells executives:

“Project 7 is the answer.”

Instead, the product philosophy is:

Research-backed. AI-powered. Human-decided.

The system should make assumptions, criteria, weights, predictions, constraints and trade-offs visible.

A decision maker should be able to ask:

Why did ProjectML recommend this portfolio?

and receive an interpretable answer.

For example:

Recommended Portfolio: A + C + F

Why?

91% strategic alignment
84% estimated portfolio value
72% resource utilisation
acceptable risk exposure
no critical resource conflicts
Project F creates a capability dependency for Project C

The precise metrics would depend on the eventual validated model; these numbers are illustrative rather than current product claims.

Who Is ProjectML For?

1. DeepTech Companies

DeepTech projects often involve:

high development costs
long development cycles
technology uncertainty
specialised talent
significant capital requirements

ProjectML can help such organisations decide which opportunities deserve scarce resources first.

2. Technology Companies

For companies managing multiple:

products
platforms
R&D initiatives
transformation programmes
technology investments

ProjectML can support portfolio-level prioritisation.

3. Enterprise PMOs

ProjectML can become an intelligence layer for Project Management Offices.

Instead of merely reporting:

“Here are our projects.”

the PMO can help answer:

“Are we investing our resources in the right projects?”

4. R&D Organisations

Research organisations frequently face a portfolio of competing opportunities.

ProjectML can support:

technology prioritisation
research investment
innovation portfolios
capability development
technology readiness decisions.

5. Venture Capital & Investment Organisations

The same framework can potentially be adapted to evaluate:

startup portfolios
technology opportunities
innovation investments
capital allocation scenarios.

This would be a later expansion rather than the initial product scope.

6. Government & Innovation Programmes

Government agencies and innovation programmes often have many potential projects competing for limited public resources.

ProjectML could eventually support:

innovation programme selection
technology funding
infrastructure programmes
DeepTech grants
public-sector project portfolios.

Initial Beachhead

ProjectML should not attempt to serve everyone on Day One.

The initial market should be:

DeepTech and technology-intensive organisations managing multiple projects under resource constraints.

This is strategically attractive because it aligns:

Founder expertise

27 years of Project and Portfolio Management

Validation research

An academic rigour and deep research backing

product capability

The product design by expert project manager

market problem

A new problem to address

FLUIDS positioning

into one coherent proposition.

Product Roadmap

Phase 1 — Research Foundation

2026

Literature review
Expert interviews
Identify decision parameters
Develop conceptual framework
Define evaluation criteria
Design research instruments
Develop hypotheses
Establish methodological foundation

Phase 2 — Proof of Concept

2026–27

Develop an initial working prototype capable of:

project data input
criteria configuration
project scoring
weighted evaluation
basic ranking
portfolio comparison
visualisation

The objective:

Prove the decision workflow.

Phase 3 — Intelligence Engine

2027

Introduce:

predictive models
uncertainty modelling
scenario analysis
Monte Carlo simulation
portfolio optimisation
explainable recommendations

The objective:

Move from scoring to decision intelligence.

Phase 4 — Pilot

2027

Work with a small number of organisations.

Target:

3–5 pilot organisations

Measure:

decision usefulness
model accuracy
interpretability
time saved
stakeholder acceptance
portfolio quality
recommendation stability

Phase 5 — SaaS

2027–28

Transform the validated prototype into a SaaS platform.

Potential capabilities:

organisation workspace
project repository
portfolio dashboards
scenario engine
AI decision assistant
recommendation engine
reporting
API integrations
enterprise controls