TAG Weekly Roundup #5

Editor's Note

Gergely Orosz reveals that successful companies like WhatsApp deliberately avoided Scrum whilst outperforming Scrum-focused competitors. The Prompt Report lists 58 techniques proving that AI communication isn't random art but systematic science. 

Both stories challenge the same myth: that more process equals better results.

📻News

Why Big Tech Ditched Scrum

Gergely Orosz's survey of 100+ companies reveals that Big Tech systematically avoids Scrum—not because they can't organise, but because empowered teams with clear ownership don't need heavyweight processes. WhatsApp's success over Scrum-focused Skype proves the point: execution beats methodology. Read more

Prompt Engineering Gets Academic Rigour

The Prompt Report, a collaboration between 32 researchers from OpenAI, Google, and Stanford, systematically catalogues 58 prompting techniques across 1,500+ papers. This isn't another "10 ChatGPT hacks" listicle; it's proof that effective AI communication follows predictable patterns.

The research shows structured approaches (context setting, example provision, output formatting) consistently outperform conversational attempts. Read more

🧰Tools

Netlify Drop - The Solution You Didn’t Know You Needed!

When Google Drive fails at webpage hosting, Netlify Drop provides instant deployment through drag-and-drop. Zero configuration, automatic HTTPS, custom domains available.

Good for prototyping stakeholder demos, hosting interactive requirements mockups for stakeholder review, and creating quick prototype demonstrations for process validation. 

Perfect for showcasing workflow visualisations to business users who need visual context rather than technical specifications. Just not your confidential requirements!

📽️Video

MCP Revolutionises AI Integration - Building on Explicit Programming’s End

This 7-minute video shows how Model Context Protocol standardises AI-tool connections. Instead of building bespoke adapters for every AI-system connection, MCP creates universal interfaces—"USB-C for AI agents."

The strategic implications run deeper—MCP-enabled tools could automate requirements traceability across multiple systems, maintain real-time stakeholder communication workflows, and provide automated impact analysis when business rules change.

It’s not about replacing BA judgement, but creating infrastructure that handles the mechanical aspects of requirements management. Imagine AI agents that automatically update user stories when regulatory requirements change, or systems that flag potential scope creep before stakeholders even raise concerns. 

Prompting 101 - The Science Behind Effective AI Communication

Anthropic's tutorial demonstrates systematic prompt engineering through a Swedish car insurance claim analysis.

The video shows how providing clear context, structured data, and explicit instructions gets reliable AI outputs.

Imagine it another way, your ability to break down complex business problems into clear, actionable components is exactly what effective prompt engineering is. The same obsessiveness you apply to eliciting requirements from stakeholders, works perfectly for eliciting useful responses from AI systems.

Till next week.