Rebuilding Our Engineering Workflow Using AI Prompts

What if every engineering project started with clarity, structure, and automation?At Singularity, we're building a new kind of workflow — one that blends engineering logic with AI precision.

🔹 What We're BuildingThe goal: to create a traceable, prompt-driven workflow that makes project delivery faster, more consistent, and easier to review.Here’s how it works right now (and yes, it’s still evolving):

1. Project Setup

   • We use AI prompts to generate the Scope of Work and Deliverables List.

   • Each deliverable has its own folder, using a naming system that includes revision control.

   • The structure is built at the start so the entire team stays aligned.

2. Context Engineering

   • We collect all specs, drawings, templates, and past calcs into a single “AI Context File.”

   • This is uploaded into the LLM, which ensures every AI prompt uses the same reference base.

3. Template Generation

   • Calculation templates are created entirely from prompts.

   • Each one includes self-checking logic for formatting and completeness.

   • We still review every template before approving it for project use.

4. Deliverable Compilation

   • Once the template is approved, we input project data, highlight the relevant scope drawing (typically a P&ID), and export the deliverable as a PDF.

   • Each final package includes a cover sheet with: Overview, Inputs & Assumptions, Results, and Methodology.

5. QA/QC Review

   • Quality assurance includes:• AI self-check• Internal Level 1 review• Final QA/QC sign-off

     
     

🔹 Strengths So Far

• The structure keeps the team aligned and traceable.

• Prompts speed up repetitive documentation tasks.

• Every deliverable is produced in a repeatable format.

  
  

🔹 Limitations (And Why It’s Still in Progress)

• Templates sometimes break formatting.

• Prompts can miss context or introduce inconsistencies.

• We’re in a constant feedback loop — updating prompts and workflows as we go.

This isn’t a finished system. It’s a live experiment. But every project we run through it helps us refine the process.

🔹 Want to Learn More? If you're building engineering workflows with AI — or want to start — our course can help.🎓 Advance your AI skills with our Udemy course: https://www.udemy.com/course/singularity-ai-for-engineers/?referralCode=75D71AF4C0EADB8975FF