Energy
Worley Deploys NVIDIA NIM Microservices to Transform EPC Business With Agentic AI
Worley
Objective
As a leading global professional services company of energy, chemicals, and resources experts, Worley was looking to unlock the potential of over 50 years of engineering expertise stored across more than 100,000 documents. Despite this wealth of knowledge, accessing the right information at the right time became increasingly difficult for their teams. To improve productivity, Worley set out to build a unified, intelligent AI platform from the ground up. This new solution, Worley AI.Assist, brings together all of Worley’s engineering knowledge and best practices into a single, scalable virtual assistant—accelerated by NVIDIA’s full-stack AI platform.
Customer
Worley
Use Case
Generative AI / LLMs
Key Takeaways
The Need for Streamlined Productivity and Collaboration
As a leading global professional services company of energy, chemicals, and resources experts, Worley is looking to unlock the full potential of over 50 years of engineering expertise stored across more than 100,000 documents. Despite this wealth of knowledge, accessing the right information at the right time has become increasingly difficult for its teams.
Determined to empower its workforce and streamline operations, Worley set out to transform the way its employees interact with critical engineering knowledge. The company’s vision was clear: Create a unified, intelligent assistant that could put decades of best practices and technical guidance at every engineer’s fingertips—improving efficiency and productivity.
Today’s engineering and design organizations, as well as asset developers and operators, manage a lot of technical content like computer-aided design (CAD) models, vendor datasheets, and compliance reports.
And with the rise of complex system interdependencies, both owner and contractor teams spend a lot of time finding the right documents, fixing version conflicts, and converting between different formats.
Within this environment, there are four main issues that affect productivity, collaboration and cost control across the industry, especially for customers seeking to accelerate delivery while maintaining quality:
- Time-Intensive Information Retrieval
- Fragmented Toolchains
- No Unified Access Layer
- Escalating IT and Support Overhead
Engineers can routinely lose 20–30% of their working hours navigating fragmented repositories and manually extracting critical data from massive document stores.
Applications for CAD, drawing comparison, simulation and quality checks excel individually—but often operate in isolation. Exporting, importing and cross‐referencing between these siloed tools creates hand-offs and duplication.
Without a single conversational or search-driven interface, knowledge remains locked in departmental vaults and application silos. Teams juggle multiple logins, interfaces and file formats to get simple answers.
Maintaining disparate engineering systems demands continual patching, API updates and vendor management. These activities increase operational spend, introduce schedule risks and divert IT resources from strategic initiatives.
These challenges left Worley searching for a smarter, scalable solution that can bring value to its customers, increase time to market, and boost productivity—one that could integrate search, analytics, and natural-language interaction across any engineering ecosystem, unify knowledge, automate routine tasks, and free up teams to focus on high-value work.
Worley
Building Worley AI.Assist
To address these challenges, Worley set out to build a unified, intelligent AI platform from the ground up. This new solution, Worley AI.Assist, brings together all of Worley’s engineering knowledge and best practices into a single, scalable virtual assistant.
What once took users more than 15 minutes—extracting information from various enterprise applications and understanding the right information—can now take less than 5 minutes with Worley AI.Assist. This dramatic improvement not only saves time, but will also empower engineers to make faster, more informed decisions.
Worley is rolling out the new platform in phases, starting with a soft launch to a few thousand users. Ultimately, the goal is to extend access to its 50,000 employees worldwide, giving everyone the ability to tap into more than 100,000 documents at the click of a button.
“Our mission is to empower our workforce with cutting-edge AI tools to drive the productivity of approximately 50,000 individuals. By integrating AI into our processes, we are not only enhancing efficiency but also providing our employees with the resources they need to excel in their roles.”
Laura Leonard
Worley Group President, Worley Digital
How Worley Integrated NVIDIA’s Platform
To realize its vision, Worley turned to NVIDIA AI Enterprise, integrating a suite of advanced technologies that now power every interaction on Worley AI.Assist.
At the heart of the platform are NVIDIA NIM™ microservices, which provide secure, production-ready environments for deploying generative AI models at scale. This approach allows Worley to rapidly orchestrate and manage multiple AI models across its global infrastructure, dramatically reducing setup time and enabling seamless updates as business needs evolve.
NVIDIA Triton Inference Server™ plays a crucial role, allowing Worley AI.Assist to serve a range of large language models and embedding models simultaneously. This flexibility ensures the platform can handle complex engineering queries and deliver fast, accurate responses—even during peak demand. By leveraging Triton™, Worley is expected to achieve a significant reduction in response latency, ensuring users receive answers in real time.
For high-precision search and contextual understanding, Worley relies on NVIDIA NeMo™ Retriever. This technology transforms both user queries and knowledge base documents into rich vector embeddings, enabling the system to quickly identify and prioritize the most relevant information. The result is a highly accurate, context-aware assistant that engineers can trust for critical project decisions.
Supporting this intelligence is a GPU-accelerated vector database, Milvus, which enables lightning-fast similarity search across more than 100,000 document embeddings. Combined, these NVIDIA technologies empower Worley AI.Assist to support real-time, multi-user collaboration for approximately 50,000 employees, all within a secure, scalable, and modular environment.
By building on NVIDIA AI Enterprise, Worley has created a future-ready foundation for innovation. This partnership not only accelerates project delivery and operational efficiency, but also safeguards data privacy and intellectual property—key priorities for a global leader in engineering and consulting.
Worley
How Worley AI.Assist Transforms Workflows
Worley AI.Assist has been built to act as an intelligent partner for engineers—to dynamically adapt to their needs through four intuitive interactions.
With conversational search, when employees ask questions, the platform can instantly scan Worley’s entire knowledge base using NVIDIA’s AI stack. It will identify the most relevant documents, then generate precise answers with source links—ultimately turning hours of manual searching into seconds of insight.
Engineering drawing or quality document analysis will enable an engineer to upload a technical file, and then the system will automatically route it to specialized tools. This ensures tasks like drawing validation or quality checks happen instantly, only when the file and query match perfectly.
With uploaded document Q&A, when users attach a file and ask a related question, the assistant will create a temporary “knowledge snapshot” of that document. Answers will be drawn from both the uploaded content and Worley’s central repository, providing comprehensive context.
And hybrid guidance will help with complex queries, from document uploads and broader questions. The assistant balances insights from the file and Worley’s institutional knowledge to deliver unified answers that reflect the full scope of expertise.
Three scalable pipelines power this adaptability:
- Agentic Workflow: Documents requiring analysis (e.g., drawings) are routed to specialized tools, with results stored securely in Microsoft Azure.
- Intelligent Retrieval: Questions trigger multi-query generation for broader search, semantic matching via Milvus, and precision reranking—all powered by NVIDIA NeMo Retriever.
- Complex Extraction: For intricate PDFs or scanned drawings, multimodal AI interprets layouts, chunks data logically, and indexes it for instant access.
Figure 1. Worley AI.Assist uses an agentic AI framework for users to interactively locate engineering tools, documentation, drawings, and diagrams for EPC applications.
Worley AI.Assist Enhances Productivity, Scalability, and Innovation
The launch of Worley AI.Assist is expected to deliver measurable impact across the organization, transforming both the day-to-day experience for employees and the company’s ability to scale innovation.
Engineers who currently spend over 15 minutes searching for and analyzing information will be able to find what they need in just a few minutes. This unified assistant will significantly reduce manual effort, freeing up employees to focus on higher-value work and drive a marked increase in overall productivity.
Worley AI.Assist is built to serve the entire global workforce, supporting approximately 50,000 users and providing real-time access to over 100,000 documents. Even during peak demand, the platform is expected to deliver reliable performance, maintaining 99.99% uptime and supporting hundreds of concurrent users without interruption.
By leveraging NVIDIA’s advanced AI infrastructure, Worley is expected to achieve rapid response times—even under heavy workloads. The platform’s intelligent routing and efficient inference workflows will ensure that answers are delivered quickly and consistently, enabling real-time decision-making and seamless project execution.
Perhaps most transformative, Worley AI.Assist will replace a patchwork of siloed tools with a single, intelligent interface. Engineers will access drawing reviews, knowledge Q&A, and document analysis all in one place—streamlining workflows and standardizing processes across the enterprise.
Worley’s investment in AI is not just about efficiency—it’s about shaping the future of project delivery to help companies across the energy, chemicals and resources sectors transition toward a more sustainable future. By combining decades of expertise with cutting-edge technology, the company is setting new industry benchmarks for productivity, sustainability, and customer value.
The success of Worley AI.Assist demonstrates how a unified, human-centric AI platform can modernize legacy systems, accelerate innovation, and empower teams to deliver on the most ambitious projects—today and into the future.
Worley
AI is advancing autonomous operations across the energy sector, including industrial digital twins, physics-based simulations, and process optimization.