Skip to content
aiqbench takeAI Products

aiqbench analysis

This article details how Saint-Gobain utilized Microsoft Discovery, an AI-powered agentic platform, to rapidly synthesize millions of patents and abstracts for developing lower-carbon cement additives. The primary focus is on the application of generative AI to accelerate materials science research and solve specific industrial challenges in construction chemistry.
Source coveringMicrosoft
What this isA model's reading of the article and related coverage, written for aiqbench. Not the publisher's words.

A concrete vision: Saint-Gobain and Microsoft Discovery build for the future

Microsoft7 min read

The post A concrete vision: Saint-Gobain and Microsoft Discovery build for the future appeared first on Source.

PARIS — Saint-Gobain’s business is construction materials. It got its start in 1665, when King Louis XIV charged the company with making the mirrors for his lavish new palace in Versailles. And he needed a lot of mirrors.

The idea was to break the monopoly that glass-making power Venice had at the time, cutting costs and improving efficiency.

These days, Saint-Gobain has 161,000 employees in 81 countries making construction materials — glass is still important, but there’s also gypsum for plaster board, insulation and concrete admixtures and cement additives.

These last elements — products for cement and concrete — represent one of the company’s most important product groups, and like glass in 1665, one of the most challenging. Cement production uses a lot of energy, and accounts for around 7% of annual CO2 emissions globally. Saint-Gobain is looking for chemical elements that could reduce those carbon emissions and improve the strength of cement.

“It would have taken our researchers a year to create the knowledge base. Using AI, it took us a couple of hours.”

Mojdeh Rastgoo-Lemaitre, a digital acceleration program manager at Saint-Gobain, said that using Microsoft Discovery saved months of time for researchers working to improve an additive for cement. Photo by Chris Welsch for Microsoft.

So, when the company’s researchers collaborated with engineers from Microsoft earlier this year in using Microsoft Discovery, a new AI-powered agentic platform for scientists and engineers, to find solutions, the results stunned them.

“We actually gathered 62,000 patents and 44,000 abstracts,” says Mojdeh Rastgoo-Lemaitre, a digital acceleration program manager overseeing the project to find lower-carbon, less expensive options. “Without using this technology, it would have taken our researchers a year to create the knowledge base. Using AI, it took us a couple of hours.”

With a broad smile, she says Microsoft Discovery helped the team to narrow the data, analyze some novel possibilities in the chemistry involved and come up with several hypotheses about new ingredients for cement additives worth testing in about six weeks.

“So, yes, the researchers are very happy with Discovery,” Rastgoo-Lemaitre says. “It’s a great tool to help them have new ideas, and to go faster.”

Saint-Gobain was one of Microsoft’s first clients to use Discovery in private release in March of this year, and teams from both companies worked on fine-tuning the platform.

The team from Microsoft was from a new operating business called the Microsoft Frontier Company, which places Microsoft engineers inside companies like Saint-Gobain to deliver solutions using AI applications and tools like Microsoft Discovery.

Discovery lets scientists orchestrate AI agents to search literature, connect information, generate hypotheses, design experiments and run analyses.

“The construction industry, from production to running, is 40% of the world’s carbon footprint,” Ursula Soritsch-Renier, chief digital and information officer at Saint-Gobain says. “We’re not going to stop constructing, but we can make change, and who is better placed to do that than us?”

She says Saint-Gobain is embracing technology as a means to cut waste, reduce energy use and be more efficient. The efforts to decarbonize the making of cement is just one illustration, she says.

“You don’t just push technology for the sake of technology. You need to always understand the ‘why’ and the ‘what for’.”

Ursula Soritsch-Reiner, chief digital and information officer, at Saint-Gobain’s headquarters in Paris. Photo by Chris Welsch for Microsoft.

For example, Saint-Gobain has changed the way it delivers plaster board to construction sites. Before, it came in a standard size, and workers cut the board to fit the rooms being built. But the leftovers just went to the dump.

“We have now a cut-to-order system,” Soritsch-Renier says. “So, we get the building information model, we optimize the cuts, then nothing goes to the landfill.”

She says the new system improved time-to-installation by 30%. “In construction, time is money,” she says. “And the environment wins because in the end it is less raw material and less energy. And that’s thanks to the digital tools we’re using.”

Microsoft has been a key partner, Soritsch-Renier says, and the framework agreement the two companies signed in July will benefit both companies by accelerating the search for sustainable construction materials and processes.

“You don’t just push technology for the sake of technology,” she says. “You need to always understand the ‘why’ and the ‘what for.’ And with that understanding you develop and you evolve and you push things forward with new tools, like with the research platform.”

Saint-Gobain’s efforts to embrace cutting-edge technology took another step at the beginning of September, when it appointed Annica Hagen to a newly created role as the company’s first chief artificial intelligence officer.

She will work with Soritsch-Renier to define how AI can support the Saint-Gobain Group’s strategic plan.

“It’s important not to talk about AI as a means to an end,” Hagen says. “It’s an enablement to get where we want to go. My main task is to enable growth — and how can we use AI to enable that? Now is the time to shape the future.”

Annica Hagen is the new (and first) chief artificial intelligence officer at Saint-Gobain. Photo by Chris Welsch for Microsoft.

Rastgoo-Lemaitre says the teams at Saint-Gobain’s Innovation Hub Europe in the Parisian suburb of Aubervilliers were already thinking about other projects where they could apply Discovery, but that research and development on the cement additives would continue.

Saint-Gobain already has a cement-additive called EnviroAdd that reduces the need for clinker in cement by about 20%.

Clinker is the ingredient in cement that binds, and it’s made by super-heating limestone, which is where the production of CO2 is most significant.

By searching across domains outside of the construction industry, Rastgoo-Lemaitre says, the scientists using Discovery were able to find some molecules they otherwise would not have considered. The idea is to create a new chemical additive that could further reduce the need for clinker in cement while making it stronger.

“They are right now testing those molecules for the behaviors and chemistry they are looking for,” she says. If the potential bears up under scrutiny, the next steps would be passing the potential solution to the development team, which would test the additive in physical conditions.

“Every day that I’m using AI today, I think, if I had only had this my life would have been so much easier, and I could have done so much more.”

Silvia Mariani is the managing director of the Saint-Gobain Innovation Hub Europe in Aubervilliers, France. Photo by Chris Welsch for Microsoft.

Silvia Mariani is the managing director of the Saint-Gobain Innovation Hub Europe, where more than 400 researchers from 30 countries work to find more sustainable and profitable ways to build. Saint-Gobain has four other innovation hubs, in the United States, India, China and Brazil.

Mariani has a Ph.D. in physics and worked on glass coatings at the innovation hub when she was a researcher. She sighs when asked if AI might have changed her work in those days.

“I have very precise moment in mind my research life where I was struggling in analyzing data, an incredible amount of data,” she says. “It was just a nightmare. And every day that I’m using AI today, I think, if I had only had this my life would have been so much easier, and I could have done so much more.”

She says her priority now is getting the best tools in the hands of her researchers and making sure they know how to use them. She added that the innovation hub would be getting Microsoft Copilot for all its employees in January 2027.

She says she is very optimistic about the future of research.

“This is a key moment. I hope we’re starting a new era of wonderful discoveries. I think we have to try to take the best of it.”

Chris Welsch is a reporter and photographer based in France covering AI, Innovation and a variety of other topics for Microsoft Source EMEA. He’s recently written about AI-guided driverless cars. Welsch was a staff editor at the International New York Times in Paris and before that a senior reporter and photographer at the Minnesota Star Tribune. Follow him on LinkedIn.

Full text under the publisher's open licence · 1362 words

Related coverage