A PLM and AI Fantasy: What the Future Might Hold

OPINION: Moving from legacy PLM to AI copilots and agile, collaborative product development.

The surge of cloud, data management and AI technologies over the last 10 to 15 years opened a new era of digital transformation where companies tirelessly try to stay ahead of the curve. The growing competitiveness of manufacturing businesses and the demand to build faster, better and at lower cost, brings many questions about how to keep up and what technologies will become differentiators. By using an idealistic and futuristic example, it’s clear to see that PLM and AI will play a significant role.

AI copilots will differentiate legacy PLM and product development workflows. (Source: Bigstock.)

AI copilots will differentiate legacy PLM and product development workflows. (Source: Bigstock.)

Digital Transformation: Marketing vs. Reality

Unless you’ve lived under a rock for the last three to five years, you’ve heard about digital transformations, digital twins and digital threads. Although these are very important topics, they come with some old products and technologies being rebranded with new names. These projects sound big and important from an IT standpoint, but for engineering teams, they don’t always sound like something that can help to build the best battery, robot, car or new engineering profession. To many, they actually get in the way of engineers thinking and innovation.

In my opinion, this is very similar to what happened a decade ago with PLM—while vendors were talking about product innovation platforms, customers were more focused on how to improve the design process and decrease the number of buttons and mouse clicks to perform a specific PLM function.

IT digital transformation projects might be a very interesting thing, but when I speak to engineers about digital transformation, they are asking how new tools and technologies can help them to be smarter and develop faster.

A wave of innovation in generative AI has triggered a lot of thinking about how new intelligent methods and technologies can be applied in product development, similar to the way ChatGPT has triggered everyone to think about automatic content creation.

To me, this is an intersection between PLM and AI. But what would that engineering world look like?

To see that potential reality, let’s explore the journey from legacy PLM—responsible for CAD file management, release and change management—to a future platform where AI acts as our copilot to build, design and develop optimal Bill of Materials, supply chains, manufacturing processes and production planning.

The Genesis of Newmo: AI Copilot Revolution

Let’s start a fairy tale…

In the fictional city of Newma Valley, the visionary engineer Dr. IK developed an AI system called Newmo to change the new product development (NPD) process. Newmo was designed to work in perfect harmony with human engineers, pushing the boundaries of complex machinery design, collaborative BOM management, manufacturing and supply chain optimization.

One day, Dr. IK’s team received a challenge from a major customer: to create a new model for a complex chipmaking machine that would redefine the industry. Newmo manufactured similar machines in the past, but the requested timeline, feature complexity and requirements for this project were beyond the level they accomplished previously. Dr. IK saw this as the perfect opportunity to showcase Newmo’s full capabilities.

To begin the project, Dr. IK’s team provided Newmo with the company’s historical data on existing machine designs, configurations, BOMs and manufacturing process planning. Newmo also received access to the data from existing PLM, ERP and SCM databases, global supplier databases and global manufacturing facilities. The AI rapidly analyzed this wealth of information and identified the required BOM configurations, features, potential inefficiencies, bottlenecks and untapped potentials in the existing BOMs and processes needed to build the new product.

Newmo then introduced an innovative NPD co-pilot application. This application supported a process that allowed the AI to work closely with human engineers, dynamically updating and optimizing the entire system design as well as focusing on detailed components and assembly design changes—including the optimization of 3D geometry and feature sets. By analyzing cost and performance trade-offs, Newmo suggested alternative components, materials and suppliers that would both enhance the product’s performance and reduce its cost.

The result was a new machine named “Nexo,” which was designed to adapt and optimize production processes in real-time, using advanced sensors and AI algorithms. Newmo’s comprehensive analysis played a crucial role in the development process, suggesting design improvements, material choices and even predicting future technological changes.

The next step was to find the best location for a manufacturing facility to build the Nexo. Newmo evaluated multiple factors—labor costs, transportation expenses, supplier proximity and regional regulations. It then ranked potential manufacturing sites based on a comprehensive analysis, calculating optimal production efficiency and cost-effectiveness of the entire process.

The results were outstanding. The new development process demonstrated an increased performance of engineers using Newmo and its capabilities to complement workloads that optimize product development and planning processes. Newmo NPD co-pilot process led to the creation of a truly groundbreaking machine with a unique and optimal set of components in the shortest possible time. Newmo not only transformed the corporation’s production capabilities but also set a new standard for AI-assisted complex machinery development, manufacturing optimization and BOM management.

The success of the project allowed Dr.IK and Newma Valley to demonstrate the potential of combining human creativity with AI in the NPD process. Their achievement opened new horizons for innovation and created a new category of intelligent NPD tools. These methods changed the approach companies use to build engineering to order products with an incredible level of product development automation, supply chain optimization, BOM cost assessment and change management.

Back in Real PLM World

Let’s come back to the reality of PLM and our current manufacturing world with all its challenges. Recently, I outlined how PLM can address five key industrial challenges. These are:

  1. Manufacturing better, faster and at lower costs.
  2. The digital transformation process.
  3. Ongoing supply chain disruption.
  4. Regulation complexity and risks.
  5. Everything as a service (XaaS).

But let’s focus on one problem and how it impacts many of the above challenges: information accuracy. Think about getting wrong data just because of an incorrect Excel file or incomplete data sync. Operating with incorrect data can lead to errors in production, resource wastage and increased costs. Therefore, ensuring data accuracy and integrity is crucial for making informed decisions and maintaining up to date manufacturing.

A big challenge for all companies is to improve their decision-making processes. The last decade of B2C innovation demonstrated that companies are getting good at giving the average consumer tools and opportunities to choose what we need (driving directions, purchasing online and more). But this doesn’t translate for engineers in manufacturing settings.

Industrial companies often struggle with data silos, complex dependencies and distributed teams, which can slow down any decision-making process. To solve the problem, companies need to integrate data and streamline communication between teams. By achieving this goal, companies can achieve better decision-making, improve the overall efficiency of the company, accelerate and streamline processes, improve supply chain decisions navigate regulations and much more.

In other words, information accuracy—or inaccuracy, in this case—is a real world roadblock to the idyllic Newmo experience.

LLM and Knowledge Graphs Accelerate Design

So how do we bring this all into reality? The question of how traditional product development can be empowered by the capabilities of AI tools like large language models (LLMs), knowledge graphs (KGs) and generative design has triggered multiple discussions in the industry. One possible path is the integration of LLMs with KGs to interpret data, generate insights and make decisions quickly and accurately. Here’s a high-level view of how this integration can support the accelerated design process:

  1. Collecting Knowledge: The language model gathers information from various sources in the PLM system and feeds it to the KG—which is like a database that focuses on data relationships.
  2. Finding Information: The AI system can then quickly pull relevant information from the KG when needed.
  3. Generating Ideas: The language model can use this information to suggest new or improved design concepts.
  4. Evaluating Designs: The AI system can compare design ideas against the knowledge in the KG, guiding teams on how to improve their designs.
  5. Making Documents: The language model can automatically create design documents using the information from the KG and the design process.

While LLM models can be efficiently used to capture unstructured information, KG itself can be built from various design sources representing existing product design and real information that exists in PLM systems. The ideas and practical examples of KG implementations already exist and are done by industrial companies and researchers.

In this picture, you can get an idea how KGs can capture information and insights form multiple design sources to be used for LLM creations later. (Image: OpenBOM.)

In this picture, you can get an idea how KGs can capture information and insights form multiple design sources to be used for LLM creations later. (Image: OpenBOM.)

When both LLM and KG are connected and integrated, they bring specific design data from a company’s files. They are captured and translated into LLM models which contextualizes data and enables engineers to easily use and produce new designs, bills of materials, production plans and supply chain optimization.

Where Does AI and PLM Go from Here?

By linking AI and PLM, the industry will take engineers through an imaginative landscape filled with game-changing ideas and disruptive possibilities, examining what it might mean for businesses globally. Industrial companies are looking at how to improve the effectiveness of decisions in product development. And by further integrating LLMs with KGs and PLM, design teams can harness the power of AI to speed up their product development process.

Our quest to understand the future of PLM and AI is a journey into a realm of endless potential. As we move forward, we will discover a huge potential for new data-driven technologies. The goal to turn manufacturing companies from traditional “electronic paper” legacy PLM processes into a data-driven world of AI and knowledge is very promising.