Impressions of INCOSE IS2025

The INCOSE International Symposium (IS) 2025, held in Ottawa from July 26 to 31, gathered around 600 in-person attendees and numerous online participants from across the global systems engineering community. The program featured roughly 160 sessions, ranging from tutorials and paper presentations to vendor tracks.

(Hot) Topics covered at INCOSE IS2025

Key topics spanned AI for Systems Engineering (AI4SE, 21 sessions), Systems Engineering (SE) theory and practice (50 sessions), Model-Based Systems Engineering (MBSE, 21 sessions), Digital Engineering (16 sessions), and SysML V2 (11 sessions). In addition, industry-focused sessions highlighted applications in defense, space, automotive, and energy.

Presentations explored AI’s role in automating SE processes, its implications in complex systems, and best practices for leveraging large language models (LLMs). Vendors showcased their

This year, all eyes were undoubtedly on AI. So let's go over all of this in this new summary.

1) AI

The keynote address of the event was delivered by Langdon Morris, a strategist and futurist who has published several books on the topic, including The AI Nation and The AI Future. Langdon presented an AI roadmap spanning from 2023 to 2031, beginning with chatbots, progressing through agents, and eventually to robots. By that time, humanity’s capabilities are projected to be 50 times greater than they are today. According to Morris, Artificial General Intelligence (AGI) could emerge by 2029, followed by what he calls “the singularity” in 2030–2031, a phenomenon beyond human comprehension.

The AI presentations at the event were grouped into three main categories:

• Using AI to automate and accelerate systems engineering processes (AI4SE).
• The implications of applying AI in complex systems.
• Best practices for leveraging large language models (LLMs).
  
  

2) AI4SE (Artificial Intelligence for Systems Engineering)

For the “AI4SE” track, several topics were presented:

• Architecture Copilot – a study on using AI to guide on system architectural issues.
• LLM for Use Case Generation – exploring how large language models can assist in creating system use cases.
• AI for MBSE and Doc to MBSE – papers focused on leveraging LLMs to accelerate Model-Based Systems Engineering (MBSE) by generating MBSE models from informal text descriptions.

One vendor, Ketryx, showcased a complete infrastructure of AI agents designed to perform various systems engineering tasks, including traceability, requirements conflict detection, and impact analysis. Another vendor, Trace Space, introduced what they call “the first AI-native requirements management platform”, which embeds artificial intelligence directly into the engineering workflow to automate traceability, change impact analysis, duplication detection, and standards compliance.

COMAC (China Aviation Company) presented research on a framework for applying LLMs to realize Pattern-Based Systems Engineering (PBSE). This approach leverages existing MBSE patterns and system data to generate new system models. Intuitive Research shared how natural language processing (NLP) and LLMs can enhance MBSE by automating tasks, accelerating processes, and reducing errors.

3) SE4AI (Systems Engineering for Artificial Intelligence)

There were also a couple of talks in the category “SE4AI” that discussed the implications of using AI components in highly regulated environments. A team from CMU discussed the AI models lifecycle and how to address the various vulnerabilities across these lifecycles. The Ansys team addressed the challenges of integrating AI algorithms into safety-critical systems and highlighted emerging standards in this area, including ED-324/ARP6983 for aeronautics and ISO/PAS-8800 for automotive applications.

INCOSE IS2025 AI hands-on workshops

On the weekend preceding the event, I attended two workshops delivered by Dr. Ray Madachy from the US Naval Academy. These sessions focused on using Jupiter Notebook with Python for two purposes:

1. Using the open-source systems engineering library se-lib to create system simulation models and perform various system analysis tasks. This tutorial is available here: INCOSE IS 2025 Tutorial — INCOSE IS 2025 Tutorial 1.0 documentation
2. Using Jupiter lab to interact with LLMs. This tutorial is available here: INCOSE IS 2025 Tutorial – Developing Custom LLMs for Systems Engineering.ipynb - Colab

This workshop covered creating custom GPTs, building Retrieval-Augmented Generation (RAG) pipelines, developing evaluation datasets, and benchmarking LLMs. Overall, it provided valuable insights into programmatic interaction with LLMs to streamline key tasks and enhance their practical utilization.

SodiusWillert at INCOSE IS2025

As part of a presentation, we showcased our three interoperability solutions: the Publisher suite of plug-ins, OSLC Connect, and SECollab. In our presentation, we highlighted how Publisher supports a wide range of MBSE interoperability scenarios across three leading MBSE tools: Cameo (Dassault), IBM Rhapsody, and Sparx EA. Our use cases included: 

• System-to-Software modeling across different tools
• Tool-to-tool migration
• Co-development among integrators using diverse MBSE environments

If you'd like to know more about any of these topics, don't hesitate to contact our team anytime.

Key takeaways from OEM vendors

Approximately 25 other commercial vendors took part in the event. Below is a brief overview of each vendor.

1) MBSE Vendors

Dassault

Dassault delivered numerous presentations, both as sponsored sessions and program papers, including a half-day UAF workshop. They presented their Version 2 of Magic Grid, along with other detailed v2 topics such as query expressions. Several presentations also focused on SoS/UAF presentations. A highlight was their Robotic ARM digital thread demo, which illustrates a path from Cameo to Catia, including the actual activation of the arm through SysML model execution.

IBM

IBM mostly featured the latest version of Rhapsody SE – their new web-based SysML V2 modeler. They demonstrated significant progress since last year, especially regarding UI’s usability and language coverage.

ANSYS

Ansys presented a digital thread vision, starting with Systems modelling with SAM – their new web-based systems architecture modeler integrated with their simulation/analysis tools. They position the solution as “open” since it also integrates with non-Ansys tools, relying on SysML v2 APIs that at least enable modular interoperability with any simulation tool that supports SysML v2.

OBEO

Obeo featured SysOn, their recent open-source SysML V2 modeling tool. They highlighted the new interoperability feature between textual and graphical notation, and also an AI assistant.

Sensmetry

Sensmetry featured SysIDE, positioned around the “System as Code” paradigm. It provides a rich textual editor based on VS Code that fully supports authoring of SysML v2 textual models. They have recently added a graphical model visualization capability. The VS Code plugin is available as open-source free software, while the premium Automato component extends functionality with execution and a code generation framework.

MathWorks

MathWorks announced a SysML v2 MATLAB API that connects with a SysML v2 MATLAB repository. This repository can interoperate with Standards-based SysML v2 modelling tools (e.g., Rhapsody SE, Cameo) for interaction with MATLAB and Simulink.

Vitech

Vitech presented their Genesys MBSE system, built around a modelling language they refer to as “CSDL” (Comprehensive System Design Language), and a methodology named “Strata” which applies a grid framework. The platform includes natural language support, along with Sidekick, a model-based review and commenting component. Vitech also claims compatibility with SysML and UAF.

Dentsu Soken

This Japan-based MBSE vendor offers a proprietary graphical modeling language, designed around diagram types commonly used by Japanese automotive customers. Their solution also incorporates risk assessment (FMEA) and supports scheduling through charts.

2) PLM vendors

PTC

PTC highlighted its end-to-end integrated engineering vision. This year, they emphasized CodeBeamer and PureVariants as their latest acquisitions. CodeBeamer now incorporates various AI-powered assistants, supporting tasks such as requirements assessments, test generation, and regulatory compliance. They also highlighted their open architecture based on OSLC.

Siemens

Siemens primarily focused on SysML v2 as a central system repository, accessible by various tools through standard SysML v2 APIs. They also demonstrated integration between Rhapsody SE and Teamcenter.

Dassault

At INCOSE, Dassault mainly featured Cameo. One presentation highlighted a digital thread workflow from Cameo to mechanical design.

3) Other tool vendors

Other vendors included:

• Jama – emphasizing the theme of “live traceability”
• LieberLieber – featuring LemonTree, a model diff and merge tool for SysML V2
• Tom Sawyer – showcasing model visualization, including SysML V2 support
• UL Solutions – presenting Stages with added AI-driven features

Summary

This year’s Canadian edition of INCOSE IS was well-attended, both in terms of the number of participants and the variety of vendors present on the exhibition floor. The main emerging theme was, to be expected, AI, particularly its application for SE, which was well-reflected in the keynotes, papers, and vendor products. Another important area of focus was MBSE, leading with SysML V2 as a new technology. It was reassuring to see that previously overhyped themes such as “Sustainability” have returned to a more balanced perspective. After all, systems engineering has always been about designing sustainable systems, whether or not the term is being promoted as a trend.