Doosan: AI at Scale, From Steel Production to Student Learning

Doosan charges ahead in the future of manufacturing and educational resources with bespoke Dataiku-powered solutions.


month project build for transformative Generative AI project

Only Minutes

required for educational content production (previously days)


Dataiku, the platform for Everyday AI, and Doosan, a multinational industrial conglomerate and the oldest company based in South Korea, have teamed up to advance the application of AI technologies (including Generative AI) in the enterprise. 

This story of innovation is an especially compelling example of the extensive power of AI to transform the enterprise and bring concrete value in many areas given the breadth and diversity of Doosan’s portfolio of holdings, affiliates, and auxiliary organizations. 

These two particular AI use cases, in very different areas of Doosan, are impressive demonstrations of Doosan’s goal to fully integrate AI into all of its business operations. The practical versus theoretical implementation of AI, including Generative AI, throughout all of Doosan is loyal to Doosan’s commitment to “AI Everywhere.” This is a focus not only on leveraging AI capabilities but also on instigating a significant cultural shift all across the company. 

Working with Dataiku empowers us to imagine the full potential of AI and Generative AI, allowing us to embed it into Doosan’s business strategies and operations.  It’s a crucial initial step in allowing us to imagine the concept of “AI Everywhere” and bring it to life, driving agile innovation that’s practical and breakthrough.

— Robert Oh, EVP, Head of Corporate Digital, Doosan Corporation

Transforming Steel Manufacturing With Dataiku

Doosan stands as one of the world’s top ten heavy equipment manufacturers. Within this legacy of industrial excellence, the head of Digital (HoD) department spearheaded initiatives to evolve Doosan into an intelligent enterprise and embark on a transformative journey with Dataiku, the platform for Everyday AI. 

The Need to Modernize Steel Production

At the Doosan Enerbility Steel Mill, Doosan’s long-established steel mill in Changwon, Korea, the volume of molten steel must be predicted to ensure that the molten steel mill that runs all day, every day, is efficiently using energy. How? Well, it wasn’t an easy task. The mill used to use scrap metal with high iron content, but now they use other additional raw materials. Due to the varied qualities of the mixed raw materials, the prediction methods needed to be quite precise, and the predictions were more challenging than before. 

The challenge was clear: Doosan needed a more precise prediction method, and they turned to AI, specifically Dataiku, for a solution.

The Electric Arc Furnace Steel Capacity Prediction Project

For 13 months, AI experts from Dataiku and Doosan analyzed the mill’s needs, crafting a customized AI solution that could adapt to the new complexities.

If we can predict the molten steel volume and prepare the right amount in the electric arc furnace in the first place, it reduces the amount of necessary refining work significantly. So, I think it’s great in all aspects, strategy-wise and time-wise.

— Dae-yong Jung, Technical Assistant Manager at Doosan Enerbility

The resulting project’s AI model predicts the steel volume with profound accuracy by working with big data from analyzing data of past work. Dataiku was able to be integrated seamlessly with the existing features of the manufacturing execution system (MES) that Doosan already employed, keeping the work all in one centralized place. 

Data insights can be translated into action at the Doosan steel mill faster than ever before with Dataiku. 

The more data we pile up in big data, the easier the job becomes because you can just insert the value pre-filled by big data.

— Kang-tae Kim, Technical Assistant Manager at Doosan Enerbility

The Unforgettable Impact of AI at the Steel Mill 

The impact that resulted from the AI solution was quick to come to fruition. Recognized in the Smarter Manufacturing category at the Dataiku Frontrunner Awards, this initiative achieved impressive accuracy in AI model predictions for steel capacity. 

In this case, we can reduce the manufacturing costs with the AI prediction system for the amount of steel tapped from an electric furnace, cutting electricity and labor costs, and reducing GHG for enhanced ESG performance.

— So-young Kwon, General Manager, Head of Digital at Doosan

It doesn’t stop there. The infusion of AI not only led to substantial cuts in energy and production costs while fostering a culture of efficiency and innovation but also played a key role in championing efforts to move towards environmental sustainability. 

Now that we can get accurate predictions, we reduced the amount of extra molten steel volume we need based on the prediction. We have new opportunities now to reduce power, carbon emissions, and so forth.

— Yong-kyun Lim, Factory Manager at Doosan Enerbility

Pioneering With GenAI: The Doopedia Project 

From the printed pages of Korea’s most beloved encyclopedia to the digital realm, Doopedia has carved out a unique space in the world of online knowledge bases. Although it’s sometimes likened to Wikipedia, Doopedia stands distinct as a structured digital encyclopedia, not a collaborative wiki, and is a reliable encyclopedia created by a group of experts in various fields. Originating from a publisher acquired by Doosan in 1985, the educational resource continues to this day. 

In a bold move to make knowledge content more accessible to the younger generation, Doosan — in collaboration with Dataiku and leveraging OpenAI’s GPT-4 — has transformed Doopedia’s intricate content system into a format that resonates with the curiosity and understanding of even elementary school students.

From Days to Minutes: Accelerating Content Creation With the LLM Mesh

Central to the transformation of Doopedia is Dataiku’s LLM Mesh, which played a crucial role in drastically reducing content creation time. With the LLM Mesh, Doosan has been able to shorten the content development process from a couple of days to a few minutes. 

This significant decrease in time has not only streamlined content production but also enhanced the interactive capabilities of Doopedia. Now, students can quickly receive comprehensive answers to queries like “Compare Tesla and Edison,” making the learning experience both efficient and dynamic. 

Doopedia managers feed articles into an LLM which regenerates the content into an output that’s better suited for a younger audience. These articles are then validated automatically to check for any hallucinations and to ensure that all the criteria and standards are met. Doopedia managers use Dataiku to review the regenerated and validated articles to edit as needed and then the content is ready for the world. 

The deployment of Dataiku’s LLM Mesh in this project underscores its capacity to revolutionize digital education, making complex knowledge more accessible and the learning process more engaging for young learners. By generating content that is more engaging and accessible, students learn and explore concepts in a way that resonates, creating real educational value.

Looking Forward With Doosan & Dataiku

The success of the Doopedia project not only highlights the potential of AI in revolutionizing traditional sectors like education and manufacturing but also sets a precedent for future collaborative innovations powered by AI and Generative AI.

Our journey with Dataiku has just begun. We are committed to exploring the full potential of AI, driving innovative products and services that deliver exponential business impacts in a rapidly evolving digital landscape.”

— Robert Oh, EVP, Head of Corporate Digital, Doosan Corporation

These use cases are a narrative of two companies coming together to challenge the status quo, to turn data into insights, and insights into value-generating action. As they look to the future, this partnership stands as a testament to the power of AI and a new era of intelligent enterprise.

The partnership between Doosan and Dataiku is a shining example of how Everyday AI and Generative AI can revolutionize traditional industries and education. Doosan’s commitment to embedding these technologies into their core processes, supported by our platform, demonstrates the incredible potential of AI to not only enhance business efficiency but also to foster a culture of innovation. This collaboration is a blueprint for the future, proving that AI can be integrated into every aspect of business operations, making it a reality for companies striving to be at the forefront of digital transformation.”

— Florian Douetteau, Co-founder & CEO of Dataiku

LG Chem: Creating Generative AI-Powered Services to Enhance Productivity

LG Chem noticed that their employees were spending a lot of time searching for safety regulations and guidelines so, with the help of Generative AI and Dataiku, they provided an AI service that helps them find that information quickly and accurately.

Read more

Solvay: Real-Time Production Cost Monitoring

Solvay uses Dataiku to monitor and improve soda ash production across 6+ plants, reducing production costs as well as energy consumption to pave the way for a sustainable business.

Learn More

SLB: Putting Data & AI to Work for Energy

SLB partners with Dataiku to drive improvements and save millions of dollars through the use of data and AI across the business.

Learn More

Toyota: Pioneering the Future
of Mobility With AI

Toyota leverages Dataiku to democratize data access, accelerate AI model development and deployment, and uncover hidden insights from vast datasets.

Learn More

Michelin: Democratizing AI for Improved Industrial Performance

Michelin uses Dataiku to democratize AI, improving quality, maintenance, machine availability, supply chain, energy consumption, and more.

Learn More