Free Registration
As part of ISBSG’s ongoing commitment to serving the IT community, registration for the conference is FREE. You will be sent a Zoom link to attend. Click on the registration button below.
Earn 4 PDUs when you attend the ISBSG conference
Attend the ISBSG IT Confidence Conference and earn 4 PDUs (Professional Development Units). PDUs enable you to earn or maintain your certification with the PMI (Project Management Institute).
This event is also valid for 4 CEPAS SCH174 credits.
Conference Program
The ISBSG conference commences at 14:00 CET. View the corresponding time zone in your region of the world on the Time and Date website.
Time: 14:00 – 14:05 CET
IT Conference 2026: An Introduction
IT Conference 2026: An IntroductionSpeakers: Julie Gledhill (ISBSG Executive Director) & Marcello Sgamma (ISBSG President)
Official conference welcome.
View Julie’s Linkedin profile. View Marcello’s Linkedin profile
Time: 14:05 – 14:35 CET
Output-Based Contracting for Agile Software Development
Speaker: Harold Van Heeringen (NESMA)
This presentation shows how ISBSG data can be used to ground output-based agile contracts in evidence rather than negotiation. Functional size provides a stable, technology-independent unit of “output”. When used alongside ISBSG’s repository of thousands of real-world projects, it becomes possible to generate industry-calibrated targets for delivery rate , throughput and quality. Drawing on experiences structuring output-based agile contracts and tenders, I will describe how to: define “output” in a way that survives changing scope and shifting teams; use ISBSG benchmark metrics to map onto contractual KPIs; set target ranges that account for the natural spread in comparable projects. Attendees will learn how to negotiate agile contracts using delivered value rather than consumed hours. They will see how ISBSG data can be used to defend the KPIs they propose — whether they sit on the client or the supplier side of the table.
View Harold’s Linkedin profile
Time: 14:35 – 15:05 CET
Software Cost Estimation: Open Problems and Research Opportunities
Speaker: Srinivasa Gopal ( Ramanujan Society for Academic Research and Promotion of Science)
Software cost estimation impacts project planning, budgeting, resource allocation, staffing, scheduling and investment. Reliable estimation is a challenge due to the dynamic nature of software projects. Estimation models that perform well in one setting may not be applicable or adaptable to another.
Key research challenges include: measuring software size and complexity, understanding productivity, calibrating models across diverse organizations and technologies, and leveraging historical data effectively. The arrival of new technologies, including AI challenge the use of traditional estimation techniques.
A promising research direction is the use of micro-level parameters that affect effort and cost during the software development lifecycle (e.g. volatility, change requests, percentage of AI-assisted code). The impact of micro-level and macro-level parameters (e.g. software size, skill level, complexity) may provide deeper insights into the conditions that influence software effort and cost.
ISBSG data helps to explore these relationships and compare traditional, statistical,
machine learning and hybrid estimation methods. Micro parameters could also allow analysis of their effect on effort at various lifecycle stages. Another direction for future work is a predictor-corrector method for software cost estimation. An initial estimate is updated as new evidence becomes available. This changes estimation from a static prediction to a continually updated process.
View Srinivasa’s Linkedin profile.
Time: 15:05 – 15:35 CET
From hours to output: Applying NPR 5333 and FSM in the mission-critical Schiphol case
Speaker: Richard Sweer ( Project Director, Finidy)
Imagine a Monday morning at Amsterdam Airport Schiphol when a mission-critical IT system goes down. Within one hour, disruption can escalate to national level; within two hours, Schiphol could potentially come to a complete standstill. This illustrates what is at stake in replacing the Central Information System Schiphol (CiSS) — described as “open-heart surgery on Schiphol.”
The Schiphol case also illustrates a wider challenge in software engineering. Agile has transformed how software is developed, but many organisations still struggle to objectively determine how much functionality is delivered, at what cost, at what speed and at what level of quality. Traditional Time & Materials contracts primarily reward effort, while management needs transparency and predictability of results.
This need resulted in NEN NPR 5333 – Output-based measurement, performance management and contracting for Agile software development and maintenance.
The presentation applies these principles to Schiphol’s NextGen AODB project, its largest IT transformation to date. Combining FSM with NPR 5333 enables planned, built and tested functionality to be objectively measured and deviations identified early. This enables FSM to connect Plan → Build → Test → Measure → Forecast
View Richard’s Linkedin profile.
Time: 15:35 – 16:05 CET
Why a Unit of Measurement Can Impact Productivity
Why a Unit of Measurement Can Impact ProductivitySpeakers: Marcello Sgamma & Luigi Buglione (DXC Technology, GUFPI-ISMA & NTTDATA)
The chosen Unit of Measure (UoM) fundamentally influences productivity estimations and the interpretation of benchmark results for a software project. Productivity expressed in Function Points, COSMIC Function Points, Story Points, Use Case Points or Lines of Codemay lead to significantly different conclusions, for the same project.
Using examples drawn from industry benchmarking practices, this presentation explores how different sizing approaches capture different aspects of software work and why mixing incompatible units can create misleading comparisons. The discussion will highlight the importance of distinguishing between product-related and project-related effort, understanding the scope covered by a measurement method, when conducting benchmarking analyses. ISBSG repositories provide a unique opportunity to investigate these effects thanks to their rich collection of projects measured with multiple sizing methods and associated performance data.
The presentation proposes practical guidelines for selecting, interpreting, and normalizing Units of Measure. It helps practitioners improve the reliability of productivity benchmarks, estimation models, and organizational performance assessments.
View Luigi’s Linkedin profile. View Marcello’s Linkedin profile.
Time: 16:05 – 16:35 CET
Estimating AI-Enabled Systems
Speaker: Arlene Minkiewicz (Unison Global)
Artificial Intelligence (AI) and Machine Learning (ML) are reshaping software systems, yet traditional software cost estimation methods fail to capture the unique characteristics of AI development. Since the fidelity of early cost estimates is important to program success, it is necessary to re-examine how estimators think about size, effort, and lifecycle structure for AI-intensive work.
This paper presents ongoing research that decomposes AI development into observable, quantifiable lifecycle activities and defines candidate size proxies and cost drivers associated with each. By combining analytical structure with a discipline of measurable, causally grounded proxies, this work advances the profession’s ability to plan and justify investment in AI enabled capability with greater transparency and consistency, independent of any specific parametric implementation. This work is intended to result in a set of cost estimating relationships that will establish a foundation for the data collection and empirical calibration needed to mature AI cost estimation into a repeatable, evidence based discipline.
View Arlene’s Linkedin profile.
Time: 16:35 – 17:05 CET
Agentic Engineering Impact on SDLC
Agentic Engineering Impact on SDLCSpeaker: Brad Clark, PhD, Galorath Inc
This presentation examines how agentic engineering — the use of AI-driven coding agents across the Software Development Life Cycle (SDLC) — is reshaping staffing, effort distribution, and role definitions in software engineering organizations.
Drawing on Galorath’s SEER-SEM effort-loading framework and a synthesis of current industry and academic sources (including Google’s vibe-coding whitepaper, Anthropic’s productivity research, GitHub/Keystone’s Copilot study, and Dr. Ray Madachy’s proposed maturity model for generative AI), the analysis models a shift from a traditional 10,000-hour SDLC effort baseline estimate to roughly 7,132 hours under agentic engineering — a ~28.7% reduction in human effort.
The reduction is uneven: Software Developers absorb roughly two-thirds of the total hours saved, concentrated in coding and integration/testing, while Software Architects see a smaller proportional decline, since their judgment-driven work in preliminary and detailed design remains largely non-consignable to AI.
View Brad’s Linkedin profile
Time: 17:05 – 17:35 CET
Human and Artificial Biases and Debiasing techniques in Software Estimation: a case study
Speaker: Roberto Meli (DPO, Italy)
Software estimation is central to planning and executing development projects, yet both human estimators and AI-based systems may have biases that undermine accuracy. Estimation can be direct, an intuitive, analogy-driven process largely governed by Kahneman’s error-prone System 1, or indirect, relying on formalized cost models. In either case it should be a statistically grounded in order to allocate resources.
This presentation maps the biases affecting estimation, addressing both the human side and artificial intelligence perspectives. The session then reviews effective debiasing methods. It also argues that extreme precision in sizing is not always needed and favors lighter, less intrusive methods such as Simple Function Points (SFP). An exploratory experiment complements the theory by comparing human estimates with those produced by Large Language Models on a real-life case. It tests whether documented human biases are replicated or transformed, and whether LLMs act as debiasing tools or as new sources of distortion. Finally, the contribution proposes a conscious hybrid approach that combines the best human practices with AI’s predictive power to improve the reliability and fairness of estimates.
View Roberto’s Linkedin profile.
Time: 17:35 – 18:00 CET