Writing Less, Owning More: The Shift in Technical Writing
summary
Describes a tested process at Deutsche Bank that turns ad hoc prompt use into a reproducible system: a shared prompt repository stored in Git, run via editor integrations (GitHub Copilot, IntelliJ/VS Code) against the source code to produce consistent first drafts. Prompts are versioned, reviewed prompts added by pull request, tested across multiple repositories, and mapped to persona-based prompts (developer, product owner, new joiner, SL2/SL3). Outcomes: fewer duplicate efforts, predictable outputs, and a lightweight review workflow that keeps humans owning quality.
AI-written summary · about
chapters (21)
- 0:00Conference context
- 1:07Personal journey
- 2:35Audience poll on AI impact
- 3:19Industry trends in tech writing
- 4:27Developers automate docs
- 5:10AI pressure on writers
- 7:16Problems with prompt engineering
- 8:19AI pressure universal challenge
- 9:02Shared prompt repository
- 10:11Prompt repo with docs as code
- 11:58Persona-based docs architecture
- 14:29Live demo of prompt repo
- 16:59Shift to amplification with AI
- 18:46Reclaiming ownership
- 20:13Future outlook
- 21:16Q&A on AI reliability
- 23:24Defining personas
- 24:47Code comments requirements
- 26:13Regulatory considerations
- 26:56Style guides design
- 28:19Closing remarks
related talks
33mStop Looking for the Perfect Prompt: The Design-First Workflow for Coding Agents
29mSupercharge Your Productivity: Mastering Prompt Engineering
44mSupercharge Your Productivity Mastering Prompt Engineering
31mHow prompt engineering projects change R&D cycle & team work in enterprise
11mFrom prompt to production
03mInterview
35mFrom Code to Clarity: Collaboration with Devels when Writing Product Documentation
45mTest Driven Prompt Programming: How AI is Changing the Way We Code