O3sigma Leverages AI to Optimize Factories and Accelerate New Client Revenue

Over seven million factories around the world work hard to produce the goods that help power the global economy and society as we know it. Yet all of them suffer from different degrees of loss and inefficiency due to downtime, process waste or poorly-calibrated machinery. The associated costs and environmental impact of this can be huge. 

O3sigma is implemented in more than 40 factories across the globe, leveraging extensive data and the latest AI models to improve efficiency and reduce waste. Its parent company,  Obeikan, is one of the largest manufacturers in the Middle East and initially developed the solution to reduce its own losses and optimize processes.

Inspired by its own success, Obeikan began to offer its AI-powered, lean factory software to other countries in the region. And after helping those factories centralize and leverage their data with Snowflake’s AI Data Cloud, it decided to scale O3sigma to other regions around the world. 

The company is now a fully-fledged AI startup, combining the latest in AI model development, extensive manufacturing expertise and institutional knowledge to help factories reduce waste, improve manufacturing floor efficiency and boost revenue.

“There’s no such thing as ChatGPT for manufacturing just yet,” says Tarik Taman, General Manager at O3sigma. “Our aim is to change that and deliver the first commercially available industrial foundation model.” But breaking new ground and getting to market quickly demands extensive AI expertise and a powerful data foundation. 

That’s why O3sigma chose to collaborate with Snowflake, running on Azure, to build a native app on the AI Data Cloud and reach the broadest possible global manufacturing audience.

A contextual, composite AI model holds huge promise for factories worldwide by providing an interconnected view of the entire manufacturing process, instead of siloed slices limited to a single task or machine.

Manufacturers rely on hundreds — even thousands — of machines in a single factory, and each one includes hundreds of parameters and components that impact efficiency and uptime. 

But understanding and optimizing all of these components manually is impossible. Even with modern digital tools, maximizing every machine is often out of reach.

Many manufacturing solution providers have noticed this gap. Now the race is on to tap into this extensive market and deliver the first usable industrial foundation model at scale. O3sigma is more than up for the challenge and knew it would need the right technology and go-to-market partner to maximize speed and agility and compete globally. That’s why it chose Snowflake.

“Right from the start, our objective was to find a modern data and AI partner we could grow with,” explains Taman. “I’d already seen what Snowflake could do in a previous job, so I knew it was the right partner for us at this crucial stage.”

With Snowflake, the O3sigma platform showed promise from the beginning. “From one of our earliest pilots, we saw results that gave us confidence in the insights this model could deliver,” recalls Taman. 

The pilot in question saw O3sigma apply its model to the NASA C-MAPSS data set, commonly used to evaluate the ability of machine learning models to predict the degradation of turbofan aircraft engines. In just two weeks and with minimal fine tuning, O3sigma’s model yielded world-class predictions with higher accuracy across operating conditions and failure models than results published over nearly twenty years. This out-of-the-box performance is critical for the diverse equipment in manufacturing environments.

Accessing everything it needs in the Snowflake AI Data Cloud has been the key to O3sigma’s success.

To create a single simple, integrated and trusted solution, O3sigma connected Apache Parquet tables and Microsoft Azure Blob storage to Snowflake through Apache Iceberg,  centralizing all data in a single data lake. From there, the company incorporated semantic data to add vital manufacturing context and used features from Snowflake Cortex AI and RelationalAI, another Snowflake Marketplace partner, to power natural language search and other important user-facing capabilities. 

Beyond its integrated data environment, O3sigma offers clients an open source solution, enabling customers to easily connect their manufacturing floor systems and machines without expensive integrations.

The end result is a system that can ingest hundreds of parameters from thousands of machines  and offer predictive and prescriptive maintenance suggestions in natural language. For example, O3sigma can suggest operators tweak a specific motor in a set of machines to achieve a different revolutions per minute (RPM) level, improving efficiency, output and revenue.

More than just a simple alert system, the O3sigma platform shows operators how to make changes and even identify root causes of potential issues before they impact the factory floor — and bottom line.

O3sigma has already seen success applying its solution to different customers’ manufacturing operations. In one case, the model suggested adjusting the motor on a single printing machine. Despite being a small change that may have been easily overlooked by a human operator, it improved printing speeds by 15% — increasing client revenue by tens of thousands of dollars. 

Similarly, when O3sigma began working with a manufacturer to optimize polyethylene terephthalate (PET) lines, its model suggested adjustments that led to a 30% increase in overall equipment effectiveness across four lines. 

The company also applied its AI and semantic modeling capabilities within a financial and wealth management operation. In this setting, the platform was used to power an AI business analyst that allows leadership and operations teams to ask natural-language questions such as mandate performance, advisor productivity or revenue risk, and instantly receive governed KPIs and explanations.

By connecting data across portfolios, client activity and operational workflows, the solution helped uncover hidden inefficiencies, explain performance deviations and identify early signs of revenue leakage.

Built natively on Snowflake’s AI Data Cloud using Snowflake Cortex AI and Cortex Analyst, the deployment reduced manual reporting and accelerated decision making. It also demonstrated how O3sigma’s composite AI models can deliver explainable, high-impact insights in regulated financial environments before being applied at scale to manufacturing.

With Snowflake supporting O3sigma both as a technology and as a Marketplace partner, the company hopes to bring these benefits to a much wider audience in the future. “We’re really excited about deploying our solution through Snowflake Marketplace,” says Taman. “It’s really going to open doors for us.”

The factory floor of the 21st century continues to evolve rapidly, with huge implications for O3sigma’s business. “We can see that the end destination for manufacturing is the dark factory, where factory floors can run almost autonomously,” explains Taman. “This is our ultimate challenge, and we know we can only get there by combining semantic, composite models that can connect to manufacturing technologies and conventional IT systems. This is exactly what Snowflake is helping us work toward.”

O3sigma’s goals go even further than the factory floor. The manufacturing sector accounts for approximately  a quarter of global emissions, and the company hopes that smarter operations and insight-led, AI-powered adjustments can reduce  this figure significantly.

“We hope we can help individual manufacturers and the industry as a whole become sustainable,” explains Taman. “With Snowflake at our side, we get the edge, analytics, AI tools and go-to-market partner we need to achieve these goals.”