I have been working in simulation for 20 years. Now, a vision I have had for a long time is finally becoming reality: detailed physics simulation and discrete-event simulation in one open system, accessible to the people who need the answers.
The pallet-yard model in the video above was generated fully automatically by AI. It is a concrete example of the workflow we are building with realvirtual. Discrete-event simulation and browser-side physics are currently in beta.
One system, different simulation questions
Sometimes the question is about the physical behaviour of a machine. Sometimes it is about the flow of material through an entire operation. How much can the system produce? Where does work accumulate? What happens if a resource is unavailable?
These questions need different levels of detail. My vision is to bring those methods together in an open system, so that choosing the right simulation method does not also mean starting an isolated project in another tool.
That does not mean every planning question needs detailed physics. It means being able to work at the level that the question requires, while keeping simulation connected to the wider planning process.
Access should not start with a €20,000 barrier
We want to change how people get access to simulation as well. Spending more than €20,000 and having to go through a sales partner should not be the starting point for exploring a planning question.
Our approach with realvirtual is direct access and an open system. The ability to share a model and integrate simulation into a planning system should be part of the workflow. A model has more value when colleagues and customers can actually use it.
The simulation is the customer presentation
For presenting a simulation, I believe PowerPoint has had its day. A slide captures one result. An interactive model lets you discuss the assumptions behind it and explore what happens when something changes.
The experience I want is simple: send the customer a web link, open the simulation together and use the model as the basis for the conversation. The simulation itself becomes what you present.
This also changes the handover. The customer can return to the model after the meeting instead of relying only on screenshots and a slide deck.
AI changes who can build the model
Building a simulation model has often been a project of its own. Before a team can investigate its planning question, it needs someone with the time and specialist knowledge to implement the model.
AI is changing that starting point. People who are not simulation experts can describe what they need and create a model by prompting, without commissioning a specialist to build every part of it. The fully AI-generated example in this video shows what that can look like.
Engineering judgement still matters. Someone needs to check that the assumptions, resource constraints and model behaviour match the intended operation. A model that runs is not automatically a model that answers the right question. AI changes the modelling effort; it does not make validation unnecessary.
Why this moment matters to me
After 20 years in this field, seeing these pieces come together is personal. I have wanted simulation to be easier to build, easier to share and part of the systems people already use to plan their work.
Watch the pallet-yard example above. Then think about a planning question your team has never simulated because building the model felt like too much work. That is where I see the opportunity.