Over the past few months, our intern Pelle van der Zaag conducted an in-depth research study into how AI Agents and Machine Learning can transform Anti-Money Laundering processes in the financial sector. Tested across a dataset of 9 million transaction records, the study compares four distinct AML detection approaches, from traditional rule-based systems to a hybrid model combining Machine Learning and AI Agents, and measures their performance across the metrics that matter most to compliance teams: detection recall, precision, and false positive rate.

What We Investigated

AML compliance remains one of the most resource-intensive challenges in financial services. Rule-based detection systems, the current industry standard, are static by design. They rely on predefined thresholds and logic that cannot adapt to evolving laundering patterns, and they are well known for generating high volumes of false positives that place a significant burden on compliance teams.

Our research set out to test whether a combination of Machine Learning and AI Agents could address these shortcomings in a measurable, operationally meaningful way.

The Four Approaches We Tested

The study evaluated the following detection configurations against a consistent dataset of 9 million records:

Our Key Findings

The hybrid model delivered the strongest results across all performance dimensions.

Combining CatBoost with an AI Agent achieved a detection rate of 93.7%, catching the large majority of money laundering cases in the dataset. AI agents were shown to increase detection recall by up to 8.7 percentage points compared to machine learning models operating alone. That is a significant improvement, and one we believe has real-world implications for how financial institutions should think about their AML infrastructure.

The reason the hybrid approach outperforms is that each component addresses a different dimension of the problem. CatBoost identifies statistical anomalies within structured transaction data with high efficiency. The AI Agent then applies a layer of contextual reasoning, evaluating whether a flagged transaction is genuinely suspicious based on broader behavioural context. Together, they catch more, more accurately.

The paper also includes a detailed breakdown of false positive trade-offs across all four approaches, an important dimension that raw detection figures alone do not capture.

What This Means in Practice

Our conclusion is that AI agents will significantly change AML processes at financial institutions of any size. The hybrid approach we tested does not simply improve detection metrics. It changes how human analysts engage with the compliance process. With a more precise and contextually aware system handling initial assessment, investigative capacity can be directed toward cases that genuinely warrant expert review.

We are convinced this is the direction the industry is heading, and this research gives a clear, data-backed picture of what that shift looks like in practice.

A Note of Thanks

This research was made possible in part through a collaboration facilitated by the Malaysian Dutch Business Council (MDBC). We are grateful for the connection they made, which contributed directly to the quality of this study.

Our sincere thanks also go to Pelle van der Zaag, whose dedication and rigor throughout this project resulted in a body of work we are genuinely proud of.

The conversation around ITIL (Version 5) has shifted. The initial excitement of “a new version is coming” has given way to a more pressing question that professionals and organizations are quietly asking: what do I actually do now?

As organisations accelerate digital transformation and AI adoption, the evolution of ITIL continues to shape how IT service management is designed and delivered. In our recent session, “What You Need to Know About ITIL Version 5,” participants gained a clear and structured understanding of the latest updates and their practical implications.

Here is a closer look at the key insights shared.

Why ITIL Version 5 Was Introduced

Today’s organisations operate in digital-first environments where products and services are tightly integrated. AI and automation are reshaping workflows, governance requirements are increasing, and adaptability is no longer optional.

ITIL Version 5 responds directly to this new landscape. It is designed to support modern operating models, emphasise measurable value, and enable organisations to integrate AI responsibly while maintaining trust and control.

A Stronger Focus on Digital Products and Services

A significant shift in Version 5 is its primary focus on digital products and services.

While earlier versions concentrated mainly on service management processes, the updated framework reflects how organisations now manage integrated digital offerings across their entire lifecycle. This shift aligns IT service management more closely with product-centric and value-driven strategies.

The Return of the Lifecycle Approach

One of the most notable structural updates is the reintroduction of the lifecycle concept.

The ITIL Product and Service Lifecycle Model now includes eight distinct stages:

• Discover

• Design

• Acquire

• Build

• Transition

• Operate

• Deliver

• Support

This lifecycle is embedded within the Service Value System and Operating Model, ensuring value is considered and managed from initial concept through ongoing support.

Service Relationships and Value Co-Creation

Another core theme discussed was the structured approach to service relationships.

ITIL Version 5 reinforces that value is co-created between providers and customers. It emphasises designing the full customer service journey and translating expectations into measurable service level metrics.

The service journey transforms demand into value through defined steps in the value chain, supported by 34 standardised practices.

Restructuring of ITIL Practices

The framework now organises practices into two clear domains:

General Management Practices (12)
Including strategy management, risk management, knowledge management, supplier management, and continual improvement.

Product and Service Management Practices (22)
Covering areas such as incident management, change enablement, service design, monitoring and event management, service level management, and software development and management.

This restructuring creates clearer alignment between governance, operational delivery, and value outcomes.

What This Means for Professionals

The session concluded with insights into the updated certification scheme and what the transition means for ITSM professionals.

For organisations and individuals alike, ITIL Version 5 represents more than a revision. It signals a shift toward lifecycle-driven, digitally integrated, and value-focused service management designed for today’s AI-enabled environment.

Understanding these updates is an important step for any professional seeking to remain relevant and effective in modern IT service management.

Automation has already reshaped the way organizations operate. From finance and HR to IT and customer service, Robotic Process Automation has helped teams eliminate repetitive tasks and improve efficiency.

Now, a new evolution is emerging: Agentic Automation.

In our latest YouTube video, we explore what Agentic Automation really means, how it differs from traditional RPA, and how organizations can start turning intelligent automation into measurable business value.

Watch the full video here:

RPA vs. Agentic Automation at a Glance

How Leading Platforms Are Enabling This

The video takes a practical look at how enterprise-grade platforms are making agentic capabilities accessible today, including UiPath and Microsoft Power Automate. These aren’t experimental tools; they’re production-ready platforms that organisations are already using to deploy intelligent agents at scale.

What’s particularly exciting is the convergence of LLMs, orchestration layers, and existing automation investments. You don’t have to throw out your RPA estate; agentic capabilities can be layered on top, extending its reach into decision-heavy workflows that were previously out of bounds.

Cybiant’s Role: From Concept to Value

Understanding the technology is one thing. Deploying it securely at enterprise scale is another. As an Agentic Automation Fast Track Partner, Cybiant bridges that gap.

Our Fast Track programme covers the full journey: training your teams to think in agentic terms, implementation across your existing process landscape, and managed services to keep agents performing as your environment evolves. We help you avoid the two most common pitfalls: automating the wrong things and deploying agents without adequate governance.

The goal isn’t automation for its own sake. It’s intelligent, goal-driven automation that delivers measurable business value securely, sustainably, and at speed.

What You’ll Learn in the Video

In under ten minutes, the video covers:

• A clear, jargon-free definition of Agentic Automation
• The core differences from traditional RPA and why they matter
• How UiPath and Microsoft Power Automate enable agentic capabilities today
• How Cybiant’s Fast Track approach accelerates time-to-value

Whether you’re an automation leader evaluating your next strategic move or a technology buyer trying to make sense of the vendor landscape, this video gives you the foundation to have more informed conversations.

At Cybiant, we believe that well-structured Public–Private Partnerships (PPPs) are one of the most powerful mechanisms to accelerate sustainable national development. Recently, we had the privilege of delivering our PPP training program in Tanzania, contributing to the country’s journey toward achieving the ambitions set out in the Tanzania Development Vision 2050, also known as DIRA 2050.

During the visit, Jan-Willem Middelburg met with the Deputy Attorney General of Tanzania, the Honorable Samwel M. Maneno. The discussion centered on how PPPs can serve as a strategic enabler for long-term infrastructure, innovation, and economic growth in support of DIRA 2050.

Why PPPs Matter for DIRA 2050

The Tanzania Development Vision 2050 outlines an ambitious roadmap for economic transformation, infrastructure modernization, and sustainable growth. Achieving such goals requires more than public funding alone. It demands structured collaboration between government and private sector stakeholders.

Public–Private Partnerships offer exactly that.

When properly designed, PPPs create long-term frameworks that align incentives, distribute risks appropriately, and ensure performance-based outcomes. They enable governments to leverage private sector expertise, innovation, and capital while maintaining public accountability and strategic oversight.

However, the success of PPPs depends entirely on how they are structured.

The Core Pillars of Effective PPP Structuring

During our PPP course in Tanzania, we focused on strengthening the key elements that determine whether a partnership delivers true public value:

1. Strong and Effective Contracts

Clear, enforceable contracts with well-defined performance indicators are fundamental. Without measurable outcomes and transparent obligations, partnerships risk underperformance or disputes.

2. A Clear and Result-Driven PPP Framework

A national PPP framework must be unambiguous, practical, and aligned with policy objectives. It should provide clarity on governance, risk allocation, procurement processes, and monitoring mechanisms.

3. Robust Financial Models and Payment Mechanisms

Financial sustainability is critical. Well-structured financial models ensure that projects remain bankable, transparent, and aligned with long-term public interests. Payment mechanisms must reward performance and protect public value.

4. Strong PPP Knowledge and Institutional Capacity

Above all, knowledge is the decisive factor. Governments need skilled professionals who understand legal structuring, financial modeling, risk allocation, and long-term contract management. Professionalization in these areas builds confidence among investors and strengthens public sector leadership.

Tanzania’s Commitment to Professionalization

It is encouraging to see Tanzania actively investing in strengthening its PPP capabilities. Building institutional knowledge, refining frameworks, and enhancing contract structuring are clear signals of long-term commitment.

Professional PPP governance does not happen overnight. It is a journey. By investing in education, training, and structured methodologies, Tanzania is laying the foundation for sustainable collaboration between the public and private sectors.

At Cybiant, we are proud to play a small role in supporting this transformation.

PPP as a Strategic Development Tool

When implemented correctly, PPPs create true win–win structures. Governments achieve infrastructure and service delivery goals more efficiently. Private partners operate within stable, performance-based environments. Citizens benefit from improved public services and economic growth.

As countries around the world refine their development strategies, structured PPP programs continue to prove their value as long-term enablers of national visions.

If your organization or government body is exploring how to professionalize PPP structuring, build institutional capacity, or strengthen your framework, we would be pleased to support you.

Public Private Partnerships Training Courses