How can we enhance nuclear operations by incorporating artificial intelligence in fleet management?
Fig1. NucleusForge MVP prototype dashboard.
By swiftly understanding regulatory requirements, conducting industry analysis, and working iteratively, I rapidly created a minimum viable product for an AI-powered fleet management toolkit, offering nuclear operators enhanced functionality to streamline compliance, workflows, and innovation within existing systems.
Team
- Dane Wesolko – Lead Product Designer
- Regulatory Compliance Specialists
- Subject Matter Experts – Nuclear Screeners
The Challenge
The nuclear industry has operated under stringent safety and regulatory frameworks since the mid-20th century, with core functionalities centered on plant operations, maintenance, and compliance with bodies like the IAEA and NRC. However, fragmented processes, outdated IT systems, and manual workflows have long plagued efficiency, leading to high operational overhead and delayed innovation. As of 2025, the surge in AI-driven demands—such as powering data centers and optimizing reactor performance—has amplified these issues, with tech firms increasingly leaning on nuclear for reliable energy, yet facing bottlenecks in licensing and compliance.
To address this, I set out with modernizing nuclear IT products, starting with ownership of the screening module. The goal was to integrate AI to reduce risks from broken systems, enhance regulatory adherence, and support emerging trends like small modular reactors (SMRs) and AI-accelerated licensing. A primary challenge was the urgency to innovate amid 2025’s regulatory shifts, including NRC’s nuanced AI requirements and IAEA’s safety evaluations, while ensuring no compromise to nuclear security. Delays in traditional processes could exacerbate operational costs, with industry reports indicating up to 20% downtime from inefficient workflows.
Finding Opportunity
I identified opportunities in the industry’s fragmented processes by onboarding deeply into nuclear IT, analyzing broken systems like manual ticket screening and disjointed data integration. Exploring AI use cases revealed potential for automation in compliance checks and predictive maintenance, aligning with 2025 trends where AI is streamlining nuclear licensing at facilities like Diablo Canyon. An aha moment came when using generative AI tools in my own workflows, realizing how natural language interfaces could parse regulatory prompts to auto-generate compliant forms, reducing manual errors by up to 30% based on similar industry pilots.
This led to conceptualizing NucleusForge as an AI toolkit that tokenizes assets, simulates digital twins, and orchestrates workflows, and directly addresses pain points in fleet management. Industry data from 2025, including DOE collaborations, showed AI could accelerate reactor construction and operations, providing a path to innovate while maintaining ASME NQA-1 standards.
Working Through
Initially, I assessed feasibility, conducting regulatory gap analyses per NRC’s 2025 AI framework and IAEA guidelines. Competitive analysis of tools like those at ORNL and INL revealed gaps in AI for licensing, inspiring my focus on natural language interfaces for ticket creation.
I developed low-fidelity wireframes for the workflow—from chat prompts to CRUD stages—incorporating AI toolkits for reproducible outputs like SOW requests. These underwent iterations. High-fidelity prototypes in Figma Make simulated interactions.
Fig2. User flow diagram for New Action Request Workflow, integrating AI Toolkits for automated form generation.
To devise a rollout strategy, I created a phased plan: regulatory alignment, design prototyping, development, and certification, to reduce overhead by 25%.
Creating an interactive prototype in Figma enabled me to convey the vision, with features like agentic AI for multi-step scenarios and blockchain for compliance trails. This prototype, built with black/white themes for AA compliance, demonstrated how NucleusForge could integrate with existing systems, undergoing refinements based on feedback.
Fig3. High-fidelity mockup of the AI Toolkits section, showing CRUD for customizable systems like the SOW Request Generator.
By adhering to standards like IEC 60880 and NRC 10 CFR 73.54, I crafted a comprehensive MVP concept. Key deliverables included:
- High-fidelity compositions—aligned with minimalist design.
- Interactive prototype—simulate AI-driven efficiencies.
These efforts ensured a strategic approach, paving the way for an MVP that enhances nuclear innovation.
The Results
Collaborating cross-functionally, I developed an MVP of NucleusForge, transforming fragmented nuclear IT into an AI-orchestrated platform. Aiming to reduce operational overhead by automating compliance checks and workflows, cutting downtime in simulations aligned with 2025 industry benchmarks. The goal is for Operators to gain faster time-to-market for innovations like SMR integrations, with AI toolkits enabling consistent outputs for tasks like SOW generation.
Clients streamlined processes, enhancing safety and efficiency amid AI-nuclear synergies, positioning the organization as a leader in 2025’s energy landscape. This reinforced compliance while fostering innovation, demonstrating potential for broader adoption.
Reflections
In hindsight, deeper ongoing competitive analysis could have incorporated more 2025 trends like AI for threat detection earlier. Structured user testing with nuclear operators might have refined the aha moment’s gen AI integration sooner. Integrating robust analytics from inception could have quantified benefits more empirically, ensuring features like toolkits added maximum value.