In our 2023 report on the Industrial AI (R)Evolution our playful title reflected our cautionary guidance about the limited application of Gen AI to industrial AI use cases. Our follow up quantitative analysis highlighted how leaders were widening the digital divide by embracing Gen AI as part of their Industrial AI toolbox. Almost 2 years after the introduction of OpenAI’s ChatGPT 3.5, Gen AI is revolutionizing platform engineering strategies at many industrial organizations.
Industrial organizations and their seasoned engineers understand the cost of adopting new technologies intimately. Having applied data science techniques and AI models for decades, they recognize that nothing, including Gen AI, comes without effort, investment, and trade-offs. Despite this, they remain emboldened by the immense potential Gen AI offers, especially as cloud hyperscalers continue to invest massively in infrastructure, AI models and platforms, and vendors’ armies of "industrial-grade" data scientists modernize their software with AI.
Building Industrial AI Value on Digital Transformation Foundations
As the AI Wars rage on, industrial organizations have been positioning themselves to capitalize on AI’s potential, building on their digital transformation efforts that were accelerated during the pandemic. With significant investments in physical assets, complex energy infrastructure, factories and fleets, industrial organizations cannot afford a complete overhaul of their systems. Instead, they have been rationalizing and simplifying their portfolios, doubling down on platforms infused with AI by their vendors industrial-grade data scientists. This approach promises to address skills gaps by adding intelligence to the petabytes of data generated by industrial organizations digital "systems of systems", accelerating profitable and sustainable business outcomes from AI.
In ARC’s Modernizing Industrial Software for the AI (R)Evolution Strategy Report we illustrated how common legacy software archetypes such as CRM, EAM, ERP, MES, PLM, supply chain, and even relatively recent industrial clouds are evolving their architectures and deliver new revenue streams, along with new Industrial AI value to their customers. As we discussed modernization for the AI (R)Evolution with customers and vendors alike, it became apparent that new software design patterns for the AI (R)Evolution would blend legacy, with yet to be realized cloud and hybrid cloud and edge benefits, and new AI requirements. Industrial organizations frequently report to ARC that their biggest challenge is implementing data governance that ensures data quality and security, with leaders prioritizing the development of industrial-grade Data Fabrics.
Leaders are Already Assembling Industrial-grade Data Fabrics
There's no shortage of “enterprise” Data Fabrics from vendors as diverse as AWS, Databricks, IBM, Microsoft, Oracle, Palantir, and Snowflake. However, the Data Fabrics employed by industrial organizations to address factory and industrial AI use cases involve diverse data types, structures, architectures, and deployment options required for high fidelity, real-time processing, and mission-critical safety and reliability. ARC believes that Industrial-grade Data Fabrics (IDFs) are needed to “weave” together a unified, seamless layer for data management and integration across the plethora of endpoints, systems, and platforms within an industrial environment to capitalize on the wide range of Industrial AI use cases from the “factory floor to the customer’s doorstep”.
IDFs need to provide standardized solutions and methodologies to address common data management challenges, such as interoperability, scalability, real-time data processing, security, and governance. They must enable end-to-end integration of data and AI pipelines across distributed edge and cloud environments through intelligent and automated systems, while also offering flexibility and customization to accommodate the unique requirements and legacy systems of various industrial organizations.
Industrial AI Use Cases Need an Industrial-grade Data Fabric
Assemble, Don’t Gamble
No one supplier can meet the complex data needs of an industrial organization’s end-to-end value chain. The Industrial AI (R)Evolution has created additional complexity for even the most advanced enterprise architecture teams. Industrial-grade Data Fabrics (IDFs) are essential for managing the diverse and complex data environments typical of industrial settings. I encourage readers who are not familiar with the operations technology (OT) domain and the unique challenges and requirements that industrial data presents, to read my more comprehensive blog on the challenges, and ARC’s wish list for the non-existent Industrial-grade Data Fabric.
As organizations strive to manage complex and diverse data environments with AI as the new opportunity (and risk), software vendors are responding to meet these needs. Industrial and enterprise software vendors are responding to this demand in several ways. They are leveraging hyperscalers' Data Fabrics and data pipelines. They are forming strategic alliances to combine domain expertise in enterprise-scale data management, real-time industrial operations at the edge, and AI and analytics lifecycles. Additionally, they are enhancing the openness and extensibility of knowledge graphs and vector databases within core enterprise software platforms.
While these collaborations and technological advancements are promising, Industrial AI leaders are not waiting to see if they bear fruit. Instead, they are proactively assembling their own Industrial-grade Data Fabrics by leveraging their core enterprise software platforms and augmenting them with:
-
Enterprise Data Historians: These are specialized databases designed to efficiently collect, store, and retrieve time-series data from industrial processes. They enable visualization and processing of historical data, as well as real-time data analysis, supporting advanced reporting and insights.
-
Industrial IoT Edge Platforms: These focus on seamless data aggregation, data cleansing, and contextualization from sensor enabled edge devices, enabling flow of relevant data and events to central or enterprise systems.
-
Specialized Industrial Analytics Solutions: These solutions contextualize and derive actionable real-time insights from complex industrial data sets, providing packaged and highly configurable KPI’s and dashboards for specific roles in industrial operations.
-
Hyperscalers' Data Fabrics: These data fabrics provide virtually unlimited scalability, allowing organizations to handle massive data volumes and high-throughput data processing; extensive global data center networks, ensuring low-latency access and high availability; a comprehensive suite of integrated services, including storage, compute, data pipelines, machine learning, and analytics.
-
Enterprise Data Fabrics: These data fabrics offer more advanced data integration, quality and lineage, governance and compliance; AI-driven enrichment with data engineering, data science, and business analytics; autonomous management, comprehensive security, contextual integration, and multi-cloud flexibility.
Industrial-grade Ecosystems
The demand for Industrial-grade Data Fabrics is driving significant innovation and collaboration among industrial and enterprise software vendors. By leveraging hyperscalers' data fabrics, forming strategic alliances, and enhancing the openness and extensibility of knowledge graphs and vector databases, these vendors are positioning themselves to meet the complex data needs of modern industrial organizations. The industrial ecosystem must rise to the challenge, break down traditional silos, and build robust data fabrics to fully grasp the opportunities presented by the Industrial AI (R)Evolution.
To Know more: Don’t Wait for Industrial-Grade Data Fabrics | ARC Advisory Group
Colin Masson
Director of Research
Colin is a member of ARC's enterprise software team. He has experience developing, evaluating, implementing, marketing and selling manufacturing, enterprise and supply chain edge and cloud solutions.
As a former Research Director for Manufacturing with AMR Research for 5 years and most recently 15 years working at Microsoft enabling manufacturers digital transformation, Colin is widely recognized across the global manufacturing ecosystem as an evangelist and advocate for the critical role manufacturers play in creating a more resilient and sustainable future for us all.
