The Challenge: Manual Reporting and Root Cause Work at Scale
The Pepperl+Fuchs Engineering team already relies on scripting and structured problem-solving tools to drive productivity, but a significant amount of their reporting work remained manual. The team produces around 300 evaluation reports a year, largely through manual copying and consolidation, a process that consumes considerable engineering time that could be spent on higher-value work.
Beyond reporting, the team also wanted a better way to capture and surface root cause knowledge. Junior engineers needed a way to key in a failure and be pointed toward likely root causes drawn from historical data, rather than relying solely on senior engineers’ institutional knowledge.
A generic, off-the-shelf AI course would not have addressed either of these needs. The team required training that mapped directly onto their evaluation reporting and root cause investigation workflows, not abstract AI concepts disconnected from their day-to-day work.
Why Heicoders Academy: A Curriculum Built Around Real Engineering Work
Heicoders Academy worked directly with the Pepperl+Fuchs Engineering team to understand their specific pain points before designing the workshop. The resulting curriculum was grounded entirely in the team's actual work, using their evaluation reports, productivity improvement reports, and historical failure data as the working examples throughout the day.
As Singapore’s leading AI First Skills Accelerator, with a track record spanning corporate clients across sectors from banking to aviation, Heicoders Academy was well placed to translate general-purpose AI tools into practical solutions for a manufacturing and engineering environment.
Heicoders Academy is Singapore's leading AI-First Skills Accelerator, with over 30,000 learners trained, 4,000+ five-star reviews, and 300+ corporate clients including Changi Airport Group, Marriott International, and Temasek Trust.
The Solution: A Full Day, On-Site Workshop in Two Sessions
The workshop was delivered as a single full-day (6-hour) session on site at Pepperl+Fuchs, capped for the Engineering team and structured around two core sessions:
- Session 1, Master Prompt Engineering: An introduction to Large Language Models and how they power tools like ChatGPT, followed by hands-on practice with prompt engineering techniques including shot-based prompting, chain prompting, chain-of-thought and tree-of-thought prompting, and system prompting, applied directly to Pepperl+Fuchs scenarios such as drafting productivity improvement reports and structuring evaluation report summaries. The session also covered strategies for minimising AI hallucinations through data grounding, negative constraints, and mandatory citations.
- Session 2, Agentic AI and Workflow Automation: An introduction to AI agents and what makes a system agentic, followed by a hands-on lab with Manus AI. Participants worked through scenarios covering evaluation report generation, engineering data analysis, root cause investigation, wide research and dashboard creation from historical failure data, alongside an introduction to email- and web-triggered automation with n8n.
Every exercise and demonstration in the workshop was grounded in the Engineering team's real work, so participants could move beyond generic examples and develop workflows and prompt patterns they could apply directly. Standout projects included a root cause analysis dashboard, a department cost tracker, performance analysis reports, and a resource allocation hub for deployment planning across different systems, incorporating the team's competency and skill matrix.
Results: Positive Feedback from the Engineering Team
The Engineering team responded warmly to the workshop, particularly the way complex concepts such as prompt engineering, ChatGPT, and Manus AI were made accessible and engaging. In their own words:
"It was a very well conducted and insightful session and the way the AI concepts, prompt engineering, ChatGPT and Manus AI were explained made the learning experience engaging and easy to understand. Thank you once again for the excellent training and for sharing your knowledge with us."
What's Next: Applying Agentic AI to Engineering Workflows
With a shared vocabulary around prompt engineering and agentic workflows now in place, the Pepperl+Fuchs Engineering team is well positioned to start applying these tools to their evaluation reporting and root cause analysis processes.
Heicoders Academy remains a committed partner to Pepperl+Fuchs as the team continues to build on the workshop and explore further opportunities to bring AI and automation into their engineering operations.
