Our Unique Features
Agentic AI
Allow our agentic AI to elicit information about your problem, then guide you in building
causal loop diagrams and validated, simulatable models. Or let it build the model for you,
in steps that you verify throughout the process. Once you have a model, direct the AI to
perform sensitivity analysis, calibration, or optimization. Our agentic AI excels not only
at building models, but also at explaining why systems perform the way they do, finding policies
to achieve the desired results, and comparing the impacts of different policies.
Our Socrates agent asks questions to learn about your problem
Our Merlin agent compares the impacts of different policies
AI Virtual Assistant
Harness the power of AI to speed and improve the development of coherent and complete Causal Loop Diagrams
(CLDs) and simulatable models! Our brand-new AI Virtual Assistant allows you to connect to different
LLM-based AI engines, provide prompts, and receive results as both diagrams and simulatable models.
You can then analyze the results, and refine and edit them based on continued interaction with the AI engine.
The Virtual Assistant has been organized in a way that makes it easy to experiment with different queries to
improve your results.
Loops That Matter™
Introduced in 2020, Loops That Matter brings clarity to your simulation data. It lights up the most important
feedback loops, giving you a simple visual guide to the forces behind the behavior. This immediate understanding
of model behavior allows you to focus on the correct variables to create a robust analysis of the results.
Assemblies
Assemblies™ are self-contained models that
streamline the modeling process by serving as building blocks for other models. You can move from a blank
page to a simulating model in a matter of minutes. Assemblies are now also available in a special section
of the isee Exchange™ where users can share their custom assemblies and download others to their software.
Designs
Designs enable you to create sophisticated interfaces with consistent aesthetics and color palettes.
With a selection of styles that are fully customizable, you can go from blank page to professional presentation
in no time. You can even use them to create interfaces for your Vensim® models.
Storytelling
Storytelling gently introduces the model structure behind the interface by unfurling parts of the
model. Annotations explain what structure means and partial simulation reveals the structure's
implications for the model's behavior. Try a live demo here.
isee Exchange™ and Sharing
The isee Exchange is a repository of content created by Stella users around the world. You can easily
publish simulations or upload models and share them with others at no cost. Users can view and run
these simulations from any modern browser. Paid private hosting options include additional features
such as data collection or running large models with an added engine on our servers. The Exchange
can be found at https://exchange.iseesystems.com.
Our Newest Features
Time range zoom on model graphs and tables
The time zooming controls in the Results Panel, Causal Lens™, Loops Panel, and Group Plot now
dynamically drive all graphs, tables, animations, and other output objects in your model.
This allows you to quickly focus on a specific time window of interest, while maintaining
context within the full simulation.
Easy export of LTM results
The Loops Panel now has an option to export LTM results. This is the same information that is sent
when using the support functionality of the AI Assistant[CH1.1], and can help you document model behavior.
Parallel sensitivity computation
Performance just got a major boost. Sensitivity analysis and select optimization
algorithms now run in parallel across all available CPU cores, significantly reducing
computation time. This is all done automatically, with no intervention required. For
optimizations using multiple starts, each start will run in a separate thread and in
parallel. The Differential Evolution optimization technique also runs using parallel threads.
Importing documentation, tags, and units from a file
The Equation Viewer now has the option to save the units, documentation, and tags for all variables
to a csv file, and another option to import some or all of this content into the model after it has
been modified outside of Stella.