Imagine you are the quality control manager at a large electronics manufacturer. You have received reports of a serious, recurring component issue for a newly released product, which unfortunately has led to a recall. Historically, the only solution would be to issue a full recall, which has significant financial, operational, and reputational consequences. However, as part of an industrial transformation strategy, your organization has implemented a digital thread framework to provide comprehensive visibility into your organization’s data. In a few simple clicks, you can now trace the entire production history of the defective product—from design to final assembly. The digital thread helps you to quickly identify a fault in a specific batch of components sourced from a single supplier. Armed with these insights, you can determine the exact scope of the affected products, work with the supplier to remedy the situation, and initiate an extremely precise, targeted recall. This swift, data-driven response mitigates customer inconvenience, and helps preserve the brand reputation of your company.
Over the last decade, this end-to-end view, has been the promise of digital threads in the industrial space, a holy grail of data touchpoints that provide a real time view of the entire lifecycle of a product or a specific process, from design all the way to end of life. This has largely out of reach for most industrial companies for two key reasons:
The data problem: Fragmented, siloed, and uncontextualized mountains of data across a heterogenous stack of technologies and modalities, that require prohibitive investments in data science techniques to be able to leverage for a specific use case, with little scalability.
Return on investment (ROI): Traditionally, it has been difficult to prove ROI for digital thread initiatives, partly due to the challenges presented by the data problem, and partly because of the complexity to action on insights, from cultural resistance to skills gaps, to mention a few factors.
Microsoft, alongside partners like PTCbelieve we are at the pivotal moment where digital threads are becoming an attainable reality for industrial customers due to two key innovations. First, the rise of unified data foundations that make data usable by securely sourcing it from systems like customer relationship management (CRM), product lifecycle management (PLM), enterprise resource planning (ERP) and manufacturing execution system (MES), and automating the contextualization aligned to any given standard or custom data model.
Secondly, the rise of generative AI, specifically, AI agents that reason using this unified data foundation and provide insights or take actions—unlocking thousands of use cases across the manufacturing value chain.
The role of AI agents
AI agents are sophisticated software systems designed to automate complex analyses, support decision-making, and manage various processes. They are productivity enablers who can effectively incorporate humans in the loop through the use of multi-modality. These agents are designed to pursue complex goals with a high level of autonomy and predictability, taking goal-directed actions with minimal human oversight, making contextual decisions, and dynamically adjusting plans based on changing conditions. AI agents can assist in various business processes, such as optimizing workflows, retrieving information, and automating repetitive tasks. They can operate independently, dynamically plan, orchestrate other agents, learn, and escalate tasks when necessary, however, AI agents are only as good as the data used to train the models that power them, and the current landscape of AI agents in the industrial space is domain specific, so these agents are confined to exclusively operate within the constraints of a single data domain, for example a CRM agent or an MES agent.
A leading example of domain specific agent is PTC’s Codebeamer Copilot. The Codebeamer Copilot supports software development process for complex physical products, like software-defined vehicles. Codebeamer Copilot leverages the Codebeamer data graph, for a connected and comprehensive view into the product development process. From requirements management to testing to release, the Copilot provides rapid insight into key areas of application lifecycle management (ALM). The result is automated requirements handling, enhanced quality control, and boosted productivity due to drastically reducing the time it takes for engineers to write and validate requirements.
Application Lifecycle management is just the beginning. The AI-powered digital thread provides agents with the combined knowledge of the entire manufacturing data estate, with multiple domains: removing their previous limitations confining them to one function.
Real-world applications of AI-powered digital threads
The era of AI and digital threads has arrived, and it’s delivering real value for the world’s leading manufacturers today.
Schaeffler
A manufacturer of precision mobility components faced a need to modernize data management, as its data previously took days to decode. Their goal was clear: find a scalable solution to uncover factory insights faster. An agent was implemented to allow frontline workers to immediately uncover detailed information when faced with unexpected downtime. This allows operators to get the line running again faster, reducing costly delays in production.
Bridgestone
The world’s largest tire and rubber company leverages manufacturing data solutions in Microsoft Fabric to accelerate the productivity of their frontline workforce. As a private preview customer, in collaboration with a Microsoft partner, the company uses digital thread and AI technology to address key production challenges, like yield loss. The query system solution enables frontline workers, with various levels of experience, to easily interact with their factory data, and efficiently uncover insights to improve yield, and enhance quality.
Toyota O -Beya
Toyota is leveraging AI agents to harness the collective wisdom of its engineers and accelerate innovation. At its headquarters in Toyota City, the company has developed a system named “O-Beya,” which means “big room” in Japanese. This system consists of generative AI agents that store and share internal expertise, enabling the rapid development of new vehicle models. The O-Beya system currently includes nine AI agents, such as the Vibration Agent and Fuel Consumption Agent, which collaborate to provide comprehensive answers to engineering queries. This initiative is particularly crucial as many senior engineers are retiring, and the AI agents help preserve and transfer their knowledge to the next generation. Built on Microsoft Azure OpenAI Servicethe O-Beya system enhances efficiency and reduces development time.
The road ahead
The journey to fully realizing the potential of AI-powered digital threads involves phased implementation. Starting with identifying the right use cases aligned to business goals, where AI agents can play a role. Secondly, identify if the right data is available and in the right standards for usability. Lastly, quickly proving value by implementing a set of initial use cases with a minimum viable digital thread and measuring and socializing its results. Achieving the AI-powered digital thread with the Microsoft Cloud for Manufacturing capabilities:
Azure adaptive cloud approach to source data from the edge, while supporting application modernization following cloud patterns.
Partner applications as systems of records, like PTC Windchill.
Microsoft Fabric as the unified data platform, and Manufacturing Data Solution in Fabric as the data transformation and enrichment service for manufacturing operations.
Microsoft first party manufacturing agents, like Factory Operations Agent in Azure AI Foundryto unlock high-value factory use cases.
Microsoft AI platforms like Azure AI Foundry and Microsoft Copilot Studio to support development and orchestration of custom AI agents.
Partner applications with agentic AI capabilities embedded, for example PTC ServiceMax AI.
Learn more
Microsoft Cloud for Manufacturing
Manufacture a sustainable future
Alfonso Rodriguez Lepage
Director Product Marketing, Microsoft Cloud for Manufacturing
As the Product Marketing Director for the Microsoft Cloud for Manufacturing, Alfonso Rodriguez oversees Microsoft’s marketing efforts for the manufacturing sector. He aims to help customers understand how the Microsoft Cloud and Microsoft’s partner networks can solve some of the industry’s toughest problems, and create a more sustainable future.
See more articles from this author
Source link
GIPHY App Key not set. Please check settings