Case Study: Łukasz – Building a Smarter, Stronger Welding Team with AI

Background

Łukasz is a hands-on Welding Manager at William Hughes, where steel is transformed into lasting products using a blend of skilled craftsmanship and state-of-the-art robotics. With a passion for building - not just machines but high-performing teams - he leads a busy department in an environment where grinders whirr, arcs flash, and robotic arms hum.

But more than anything, Łukasz sees his real job as building people:

“Pushing people without better tools leads to burnout, not brilliance.”
His ethos is clear: work smarter, not just harder.

The Challenge: Complexity, Downtime, and a ‘Documentation Dragon’

The workshop’s biggest pain point wasn’t just the robots breaking down - it was the chaos that followed. A robot would stop, error codes would flash, and experienced techs would scramble while newer staff stood by, unsure how to help. Valuable time was lost searching through binders, emails, or waiting for help. Knowledge was often trapped in people’s heads, making onboarding slow and mistakes too common.

On top of that, the team had to deal with:

  • Fragmented technical documentation (scattered PDFs, outdated files)
  • High pressure, tight deadlines, and complex troubleshooting
  • Communication challenges, especially for Łukasz (English is not his first language), which added friction to collaboration and reporting.
Auto-generated description: A robotic arm is welding a metal grid in an industrial setting.

Practical Applications: What’s Live and What’s Next

What’s Live and In Daily Use

AI-Powered Troubleshooting (In Full Operation)

Łukasz and his team have deployed a custom AI assistant, trained on their own manuals, logs, and workshop photos. When a robotic welding arm halts, any operator can snap a photo of the error, upload it, and get back an instant, step-by-step checklist -translated into clear, practical instructions, often with helpful diagrams.

Results:

  • Minor issues now fixed by operators in minutes
  • Complex problems summarised for the maintenance team - no more guesswork
  • Robot downtime dropped by over 70%
  • Newer team members feel empowered; veterans save time and frustration

What’s In Progress, Piloted, or Being Rolled Out

Living Knowledge Base & Smart Document Search (In Progress)

Łukasz has led a major project to digitise and centralise every relevant manual, note, and email into one searchable library. Powered by AI tools like NotebookLM, the goal is for staff to be able to ask questions about machine settings, procedures, or safety standards and get direct, referenced answers in multiple languages. This system is in the testing and gradual rollout phase, with full deployment planned soon.

AI-Assisted Communication and Decision Support (Testing Phase)

AI tools for drafting and polishing emails, reports, and live decision support are being piloted. These are designed to help with shift planning, technical queries, and everyday English communication - especially useful for a multilingual team. Early results are promising, but the tools are not yet fully in place for everyone.

Weld Program Creation and Process Automation (Proof of Concept)

AI-driven weld program suggestions from 3D models, auto-logging, and smart code templates are being explored as proof-of-concept projects, aiming to make programming faster and reduce reliance on undocumented “tribal knowledge.” These are currently in the experimental stage.

Note on Implementation:
As of July 2025, the main AI tool in active, daily use by the team is the photo-based diagnostic assistant for troubleshooting robot errors. The central knowledge base, document search, and AI-assisted communication solutions are actively being developed, tested, or rolled out, and are not yet fully live for the whole team.

Real-World Impact & Insights

Time Savings & Efficiency
Łukasz can’t put an exact number on the hours saved, but says:

“I’m definitely more productive than I was before. Things run more smoothly and get done quicker.”
He now spends less time firefighting and more time on continuous improvement.

A Culture Shift
Perhaps the biggest change is cultural: the team feels more ownership, less stress, and more pride. The AI isn’t seen as a threat, but as a helpful partner. Morale is up, and operators are more proactive.

Scalable Knowledge
The knowledge base grows with every problem solved. Expertise is no longer lost when staff leave; it’s captured and shared forever. Supervisors can now focus on improvement, not just firefighting.

Looking Ahead

Łukasz sees this as just the beginning:

“Some of the tools described here are still being rolled out or refined. Our custom diagnostic AI is already making a difference, and we’re excited to bring the rest of the vision to life in the coming months. I see a future where AI predicts issues before they happen and guides training with augmented reality. The goal? A true learning organisation where humans and machines build better things together.”

Attribution:
Case study written in collaboration with Łukasz, Welding Manager at William Hughes.

Case Studies