Article

RSInsight+ is Transforming Engineering Workflows with AI-powered Scripting

Published on: Oct 24, 2025 Updated on: Nov 26, 2025 Read: 7 minutes
Author:
  • Steve Chai, Geotechnical AI Manager at Rocscience

Engineering has always been a dialogue between human intuition and computational precision. But as models grow more complex and deadlines tighten, that balance has shifted. Engineers now spend more time learning, setting up, and iterating than interpreting results or making critical design decisions. That barrier stops many teams from adopting advanced tools altogether. Training takes time. Expertise takes years.

The November 2025 update in RSInsight+ removes this hurdle. As Rocscience’s applied AI assistant, it embeds analytical precision and speed directly into engineering workflows (currently in RS2 and RSpile); transforming time-consuming modelling into an intent-driven dialogue. Engineers no longer wrestle with command sequences or documentation searches; they simply express what they want to achieve, and RSInsight+ guides, suggests, or executes.

The outcome: faster onboarding and drastically reduced training overhead, allowing teams to focus on insight rather than procedure. This piece spotlights the program's practical use cases, a detailed case study, and key takeaways for integrating AI into everyday engineering analysis.

Note: For a limited time after launch, RSInsight+ will be available to all users and will continue to be included at no additional cost for those with a valid Rocscience program license.

The Analytical Challenge: From Documentation to Decision

How can engineers minimize time and uncertainty while optimizing design for layered soil–structure interaction or pile group performance?

Mastering advanced analysis tools has traditionally meant long hours of study, trial and error, and continuous parameter refinement. While thorough, this approach is time-intensive and can slow design decisions that demand quick, confident judgment.

RSInsight+ changes this for all levels of users — from first-time modelers learning to apply supports or define boundaries, to seasoned analysts performing sensitivity studies or optimization routines. Acting as a built-in subject-matter expert, it understands Rocscience programs deeply and guides users through model setup, calibration, and results interpretation via intuitive, natural-language prompts rather than manual trial and error.

A Paradigm Shift in Applied Analysis

Think of RSInsight+ as an analytical co-pilot; one embedded with the domain knowledge and validation logic of a veteran Rocscience user. It suggests basic modelling approaches to automate repetitive modelling steps, accelerates parametric studies, and ensures that every adjustment follows validated analytical logic, not guesswork.

By leveraging Rocscience’s solver intelligence, RSInsight+ interprets user objectives, performs actions within the model, and, when required, iterates through parameter adjustments — all within a single, guided workflow. This intent-based interaction flattens the learning curve and allows engineers to move from setup to insight faster, without compromising control or analytical rigor.

Use Case Scenarios

  • Getting Started: Assisted Model Building:
    Even simple tasks, such as applying a bolt or defining a boundary condition, can be streamlined through natural language interaction. A prompt like “Add a bolt at 3.5 m along the excavation face” leads RSInsight+ to suggest the correct input parameters and confirm before applying changes directly in the model. For new users, this is an interactive learning experience that accelerates onboarding and builds confidence from the first session.
  • Parameter Tuning for Targeted Settlement Reduction:
    Before: Analyzing and calibrating soil parameters to meet settlement targets was a lengthy and time-consuming process. It typically involved manual adjustment, multiple simulation rounds, and periodic cross-referencing with documentation.
    Now: With RSInsight+, a single prompt — e.g., “Reduce total settlement to less than 15 mm by adjusting soil stiffness” — returns actionable results in moments, supported by cited analysis steps and direct model modifications.
  • Batch Processing and Automated Result Compilation:
    Managing numerous file configurations and extracting comparative results is reduced to a near-instantaneous output, with RSInsight+ compiling results into ready-to-analyze summaries; ideal for design comparisons and report preparation.
  • Model Configuration and Data Import:
    Importing and analyzing variant geometries, often slowed by repetitive preprocessing, is reduced to seconds. Engineers can explore design alternatives or run what-if studies with minimal setup effort.
  • Optimization for Engineering Criteria:
    Complex optimization routines, such as adjusting stiffness or pre-tensioning to meet defined design goals, are simplified into a single conversational exchange; completed through focused, guided analysis. What once took hours of iteration now requires only minutes, allowing engineers to concentrate on validating outcomes and refining design strategies.

Case Study: Anchored Pile Wall Optimization with RS2

Consider an anchored pile wall, mid-design, with the objective of reducing the structural liner’s displacement at a critical stage by 20%.

The Prompt:

“Change the pre-tensioning force of bolts so that displacement of structural liner at Stage 5 is reduced by 20%.”

The Process:
RSInsight+ interprets the command, modifies bolt properties within RS2, executes the staged analysis, and reports results both in conversational output and in the updated model.

Figure 1. Initial prompt for iteration of bolt property
Figure 1. Initial prompt for iteration of bolt property
Figure 2. RSInsight+ agent modifying the model directly
Figure 2. RSInsight+ agent modifying the model directly

The Outcome:
RSInsight+ iteratively adjusts the bolt pre-tensioning force, resulting in a 20% reduction in displacement at the top of the liner. The change is validated both in the summarized chat output and visualized directly within the RS2 model results (see Figure 3).

Figure 3. Final compiled results in RSInsight+, displaying chat-based analytics (left) alongside RS2’s graphical output (right)
Figure 3. Final compiled results in RSInsight+, displaying chat-based analytics (left) alongside RS2’s graphical output (right)

Technical Reflections & Best Practices: Ensuring Analytical Integrity

  • Prompt precision matters: Clearly defined objectives and constraints yield the most reliable results. Engineers retain full oversight and can review, or override changes as needed.
  • Comprehensive traceability: Every automated change and iteration is logged, ensuring transparent tracking of all model adjustments.
  • Fallback and rescue mechanisms: If a prompt isn’t supported or constraints are exceeded, RSInsight+ flags the issue, logs it, and enables manual rollback.
  • Respect for modelling limits: For example, AI-driven iterations in RSPile obey all constitutive and structural constraints (e.g., nonlinear t-z, q-z curves).
  • Seamless workflow integration: RSInsight+ layers atop the full capabilities of RS2 and RSPile: staged excavation, seepage, consolidation, group and lateral analysis; ensuring established workflows remain intact.
  • Engineer validation remains essential: Automation accelerates iteration but does not replace engineering judgment. Final designs must always be validated against project and safety requirements.
  • Future-ready integration: The decoupled prompt engine in RSInsight+ sets the stage for future integration with the full suite of Rocscience tools.

Analytical Rigor, Empowered by AI

By codifying expert knowledge and integrating directly into the RS2 and RSPile workflows, RSInsight+ enhances both analytical speed and reliability. Engineers can perform focused explorations, validate parameter sensitivities, and mitigate sources of error — all while preserving the traceability expected in high-stakes geotechnical engineering.

Conclusion

RSInsight+ represents more than an analytical assistant; it is an accelerant for informed engineering decisions. It augments geotechnical practice by refocusing where they spend effort: from the rigmarole of manual simulation to interpretation, judgment, and design decisions. Across use cases, from bolt modelling to batch analysis, the transformation is tangible: workflows that once demanded hours now culminate in actionable understanding within minutes.

The real breakthrough lies in how RSInsight+ embeds the domain knowledge of Rocscience programs (currently RS2 and RSPile), enabling engineers to express geotechnical intent while the system executes the low-value setup and iteration. Over time, prompted workflows mature, transparency deepens, and iteration becomes fluid instead of onerous.

Blending Rocscience’s analytical legacy with applied AI intelligence, RSInsight+ defines a new era of engineering productivity, where precision meets speed and expertise is instantly within reach. Complex ground–structure problems are solved faster, with full traceability and control, placing analytical mastery squarely in the engineer’s hands.

Engineers lead the solution; RSInsight+ ensures every detail counts and no moment is wasted.

Be among the first to integrate this paradigm shift into your applied analyses!

Start Your Free Trial
Back to top

More from Rocscience