Thirty Years of Rocscience: From Geotechnical Tools to Agentic Engineering for Five Disciplines

CEO Thamer Yacoub looks back at what started as a research group at the University of Toronto and became a global software company that thrives on meeting complex engineering challenges.

Software is a small fraction of an engineering project's total cost, but it sits at the critical point where a wrong decision will be measured in the loss of lives, property, and reputations. Rocscience has spent 30 years developing tools engineers use to reduce both safety and economic risks. As the name implies, the company started with rock but has broadened its focus to serving other engineering disciplines and the challenges posed by other materials, such as concrete, steel, and water. Achieving the three-decade mark this year invites not only retrospection but also the opportunity to look at what lies ahead for engineering technology.

Origins in Academia

The Rocscience story began in 1987 at the University of Toronto, where Dr. John Curran formed the Rock Engineering Group to build 2D geomechanics software for mining and civil engineers who needed reliable tools that were easy to use. John was encouraged by Dr. Evert Hoek, who had studied rock mechanics in the gold mines of South Africa and co-created the Hoek-Brown failure criterion.

In addition to John, the group included graduate students Thamer Yacoub, Brent Corkum, and Reginald Hammah, now Rocscience’s CEO, CTO, and CSO, respectively. They wrote what would become the company’s first programs, backed by the Canadian mining industry and government partners — real funding for real software, not just academic curiosity. Thamer personally wrote the code for the boundary element engine of EX3, the finite element engine of RS2, and the groundwater engines of Slide2 and RS2. He was also the lead developer of Settle3.

To handle growing demand, John formally established Rocscience as a company in 1996, after the software had already proved itself. Growth in the early years came from working with mining customers in Australia and New Zealand, who were introduced to DIPS and other tools created by the Rock Engineering Group, with Slide2 and RS2 later becoming industry standards in slope stability. Rocscience added 3D modelling with the release of RS3 and Slide3, making the company a leader in both 2D and 3D geotechnical software before it began acquiring companies in other disciplines.

“It was a dream come true and mind-blowing to go from PhD research to commercial software serving geotechnical and mining customers who depend on our solutions in their daily workflows,” Thamer explains. “We see our products as our babies because they were developed from scratch and now they're implemented across the world to solve engineering problems,” he added. “What's even more gratifying is seeing customers push the envelope with them, expanding their uses beyond what was initially imagined.” Settle3, for example, was designed to do simple settlement analysis for foundations, but customers now use it to engineer embankments with complex, multi-stage drainage systems — evidence, in Thamer's telling, of tools that keep growing to meet the challenges people bring to them.

One Platform for Integrated Workflows

Expansion from geotechnical solutions into other disciplines began in 2024, when Rocscience acquired DIANA, bringing advanced finite element analysis for complex structures into the portfolio, and 3GSM, which added rock mass characterization for mining. In 2025, 2SI added further structural engineering capabilities and Rockfield brought wellbore stability and hydraulic fracture modelling for energy geomechanics. In December 2025, Rocscience completed its acquisition of Aquanty, adding integrated groundwater and surface water modelling to the platform. The acquired products, all highly specialized, enable Rocscience to build integrations within and across all five disciplines.

“We're creating an engineering ecosystem, a platform that allows users to move from one tool to another to support a specific workflow,” Thamer says. He gives a few examples of what that looks like in practice. In mining, a single project might touch ground stability, rock mass characterization, and dewatering across the same lifecycle — work that used to mean juggling separate tools and separate vendors, now handled together. In structural engineering, a settlement analysis can connect directly to the superstructure of the building or bridge sitting on top of it, instead of being handed off between separate teams. In energy geomechanics, a single wellbore can be modelled across its entire lifecycle, from drilling and hydraulic fracturing through completion, production, and injection, rather than treating each stage as its own disconnected analysis. And water shows up in almost every project, so Thamer describes a near-future where real-time weather, rainfall, and infiltration data feed directly into slope-stability analysis, without an engineer manually collecting each input first, freeing engineers to “focus on being engineers rather than just collecting data.”

The actual job of an engineer isn’t about spending without limit to eliminate risk; it’s finding the answer that’s safe and also economical. But that job keeps getting more difficult because materials behave differently, conditions grow more demanding, and the weather and environment are changing. Today, reaching a single slope-stability answer might require limit-equilibrium analysis in Slide2, deformation analysis in RS2, hydrogeology data from HydroGeoSphere, and seismic analysis in RSSeismic — four tools working as one system to produce the needed calculations. “Our goal is to facilitate the real-time, multi-dimensional modelling and analysis engineers need to design a safer world, while also factoring in overall value.”

Nearly half of Rocscience's team, close to 100 people, sits in R&D — a commitment to innovation that helps engineers mitigate risk and make the right decisions, not just ship software. That investment now extends to how the company builds its own tools. More than 80 percent of the code Rocscience ships today is AI-generated, Thamer says, with Rocscience's engineers reviewing and directing the work rather than writing every line by hand, cutting a development cycle that used to take a year down to roughly four months. It's a concrete example of AI already in production, not just a concept — getting customers what they need and ask for, sooner.

What Comes Next: Engineering Software That Talks Back

Customers have been asking for that same kind of automation for their work, too, although engineers may be skeptical of AI in high-stakes work, where a wrong output doesn't just look bad — it can fail. Rocscience's agentic AI platform is built around that skepticism, not in spite of it. Ask Thamer directly about the risk of handing more of that process to AI, and his answer doesn't hedge. The software is there to assist the engineer's analysis, not to take over the approval or sign-off. It still needs an engineer's input to run, produces an output someone has to interpret, and leaves the actual decision — the one that depends on “background and judgment” — with the person, not the model.

That's also the role RSInsight is meant to play as the platform keeps growing. Rather than building a one-off connection between every pair of tools, RSInsight is intended to become the common AI layer that ties the whole ecosystem together, making model setup easier and accelerating analysis, reporting, and technical decision-making. Today, it has two specific jobs — making sure data migrates correctly as engineers move between tools, and automating the repetitive parts of a project, running multiple analyses, formatting reports, “so engineers can focus on what they do best.”

That same idea extends to how engineers in the various disciplines will interact with the software itself, not replacing an engineer's judgment, but turning an engineer's own Natural Engineering Language (NEL) prompts into cross-product workflows, rather than requiring the engineer to translate their thinking into code or clicks. The idea is for AI to handle the how of analysis, so engineers can focus on the why of design.

Engineering analysis is rarely a one-time exercise — it means iterating through what-if scenarios to reduce risk, which means a lot of repetitive work along the way. What customers want, in Thamer’s words, is to “leave the tedious work of repeating things and let the scripting and AI help them,” so their own time goes toward judgment calls instead of setup.

What Hasn't Changed

Thamer reflects on the company's culture, tracing it back to what he told John early on. “I'll stay at Rocscience as long as the work is fun.” Thirty years later, it obviously still is. “I'm so honored and blessed to have an amazing team,” he says. Rocscience now has more than 200 employees and offices in Canada, Australia, Austria, Ghana, Italy, the Netherlands, Peru, and the United Kingdom, part of a deliberate plan to serve customers in their own regions and time zones, not just from a single head office. The company's connection to academia remains strong, too — Rocscience maintains more than 600 academic partnerships worldwide, offers Academic Bundles that give institutions real-world training tools for the classroom, and just welcomed its newest cohort of engineering interns for the start of the fall semester.

The company's customer-first instinct shows up in small, personal moments as much as it does in big strategic decisions. Thamer was recently in South America for an industry tradeshow and previewed a new piece of mining software the company plans to release later this fall. Watching a prospective customer realize that Rocscience is working on a solution to a problem they've been wrestling with is part of the job he says he enjoys most — “you can see the smile on the face of the user — it made my day.”

Rocscience's tools are ultimately for giving engineers solutions they can trust to do the essential work of creating designs that are both technically rigorous and economically sound. The last 30 years, and the goal for the next 30, is continuing to help engineers make confident decisions on the projects that matter most.

Asked what he wants customers and the industries Rocscience serves to know, Thamer doesn't reach for a grand statement about the future. He goes back to the same ethos that's carried the company from a university spin-off to a leader across multiple engineering disciplines. “We thrive on challenges. My email is always open, to hear about their challenges, suggestions, and cool ideas. This is what I say in every user group, it’s what I ask in every meeting: how can we help you?”