Accessibility is key as AI continues to transform manufacturing. By embedding AI steadily and responsibly into workflows using the crawl, walk, run, fly framework, businesses can deploy analytical AI tools swiftly and agilely.
Rahul Garg • VP for Industrial Machinery Vertical Software Strategy | Siemens Digital Industries Software
Workers need to take AI into their own hands to discover its full potential. Image: Siemens
The AI revolution is here and accelerating. From generative AI to large language models (LLMs), AI-powered technology has helped many manufacturers flourish in this decade, and the industry expects advancements in AI to gain momentum. Savvy corporations anticipate AI solutions to continually bring their operations to higher levels of digital value as their analytical and automation capabilities improve. Now, with the advent of agentic AI, it is poised to further transform industrial manufacturing and unlock new functionalities that will propel the industry further beyond.
Even in this age of “AI first” frameworks, machine builders and equipment and component manufacturers must still empower their people to help shape the future of design and engineering. While AI brings a host of functionalities that improve productivity, optimize operations, and increase efficiency, these benefits are maximized by keeping humans in the loop. To guarantee that workers build trust in AI and use it as an indispensable partner, organizations must first make AI accessible.
Accessibility starts with maturity
The first — and arguably most important — step to ensuring that a business keeps its people in the loop is understanding the digital maturity of the organization so that AI is deployed at the appropriate level. This whole process of transformation looks very different in many companies, and approaching AI transformation holistically sets them up to handle very complex sets of challenges in the future. Taking a structured approach to AI integration encompasses four phases:
- Crawl
- Walk
- Run
- Fly
Making AI accessible helps businesses take full advantage of their data. Image: Siemens
Starting simple and gradually introducing more complex AI functionalities will ensure that people remain in the loop and set up enterprises to adapt to each new AI solution more easily. In the crawl stage, companies should get a feel for their specific AI needs. This way, workers have time to adapt to new solutions, and the company can determine what is working. Straightforward tools are good stepping stones; command prediction during the product or part design process can reduce repetitive tasks for the engineer and ease onboarding. Even tools like AI-assisted repair for understanding error codes and identifying the root cause problems provide immeasurable value.
Once the organization is comfortable with these less demanding solutions, it’s time to move on to the walk stage, which includes introducing more advanced features such as performance predictors and AI-enhanced topology optimization for part and component design. From there, businesses can begin to run and start training AI models using company-specific best practices and historical data collected during the crawl and walk stages.
And in the fourth stage, fly, machine builders can fully integrate generative AI and agents to create new engineering and manufacturing content and automate complex workflows. Already, customers using AI-driven workflows have reported efficiency gains of 10 to 20%. Tools such as the Designcenter Copilot — a tool-path creation copilot that functions similarly to ChatGPT but is embedded within the design environment — provide real-time, context-aware suggestions. These tools help identify potential design failures early and guide users through corrective actions using machine learning models.
Training a smarter partner
AI is quickly becoming a trusted partner in engineering and manufacturing. Only last year, many manufacturers began deploying LLMs to fill in knowledge gaps and bridge human and machine operations.
LLMs have significantly accelerated the implementation of new solutions on the shop floor by transforming how operators, engineers, and systems interact with industrial data. Tools such as Insights Hub Copilot cater to the AI needs of machine operators, enabling operators to interact with shop floor data using natural language, helping anyone — regardless of skill level — make the most of factory data.
AI for everyone
Businesses should strive to deploy solutions that cater to everyone from the shop floor to the top floor. When all skill levels on the shop floor can reap the benefits of data-driven manufacturing, it ensures that organizations can efficiently and effectively use their data — regardless of the individual business’s digital maturity.
Designers and engineers can use AI to optimize their design workflows. Image: Getty Images
For example, many design engineers and production engineers do not have formal training in app development or programming, but this does not prohibit their ability to leverage AI. There are solutions now available that require no technical AI skill yet can empower operators to take full advantage of shop floor data. Designed to be simple and intuitive, these tools provide transparency into equipment status and key performance indicators while monitoring equipment and enabling predictive maintenance. Meanwhile, engineers with a deeper understanding of machine programming and production processes can use AI to configure tailor-made tools for quality prediction and process optimization.
Additionally, due to the versatility of AI-infused tools, their solutions can be leveraged by both large businesses and SMBs. SMBs benefit from the low barrier to entry due to its Software as a Service model, eliminating the need for heavy on-site infrastructure. Meanwhile, large enterprises can leverage the scalability and easily integrate the capabilities into their existing systems.
Industrial foundation models
Siemens is developing industrial foundational models for more insightful LLMs. Unlike general-purpose language foundation models, such as those used by ChatGPT, industrial foundation models will be trained on industrial data, including everything from CAD files to simulation data and validation reports. Industrial foundation models will understand the language of engineering and manufacturing, making them context-aware and able to reason with complex industrial data that is hard to put into words, such as 3D models and sensor readings.
Many companies have decades of legacy data in the form of 2D drawings and product manufacturing information (PMI). AI tools can learn from this data to support engineers to tackle time-consuming tasks more productively, ensure compliance with industry standards, and enhance and accelerate the product development process across design, planning, engineering, operation, and service. To protect IP, these models can be deployed on-premises, maintaining data privacy while enabling tailored AI solutions.
Context-aware AI can become a virtual partner. Image: Adobe Stock
Benefits
Manufacturers that adopt data-driven manufacturing approaches that are accessible for every worker see more streamlined maintenance processes, improved throughput performance, and even more efficient energy usage. Benefits some companies have seen include:
- Greater than 30% increase in asset availability via predictive maintenance.
- 50 to 60% reduction in nonconformance costs through AI-driven quality prediction.
- Up to 30% fewer defects due to improved transparency and root cause identification.
- More than 20% performance gains using IoT data in plant simulation models.
- Up to 30% energy savings
If implemented using the crawl, walk, run, fly framework, businesses can deploy analytical AI tools swiftly and agilely across even a global organization. Some manufacturers have been able to roll out data-driven solutions to 80 plants in 20 countries due to the ease of configuration and integration into already existing systems.
Getting ahead is easy
Accessibility is key as AI continues to transform the manufacturing landscape. With AI solutions becoming more advanced by the day, companies are reaping the benefits of starting their digital transformation journeys. Integrating AI is now easier than ever — and it all starts with data.
Embedding AI steadily and responsibly into workflows ensures it reflects the creativity and expertise of the engineers it supports. LLMs help unlock data on the shop floor, enabling operators to accurately analyze and monitor their data, and many forecast LLMs to become even more of a boon for manufacturers and machine builders alike as businesses continue to work on an industrial foundational model. The future is clear: companies will be using intelligent tools to drive efficiency, innovation, and transformation across every aspect of engineering.
About the author:
Rahul Garg is Vice President for Industrial Machinery Vertical Software Strategy at Siemens Digital Industries Software. As a customer-centric leader, one of his great joys is helping simplify complex problems for customers and enabling success by delivering powerful, effective solutions that support small and mid-sized businesses. Throughout his career, having worked at three start-ups and now a large enterprise, Rahul has worked closely with SMBs in technology-led industries to overcome key challenges and drive revenue growth with strategic solutions, smarter services, and better business practices. Connect with Rahul
Siemens Digital Industries Software
sw.siemens.com
Filed Under: AI Engineering Collective, AI • machine learning
