Build cluster · Product, design, research
Technology Rewired: Product and Design
Product and design tools moved from desktop design apps and requirement documents to cloud design, roadmapping, and experimentation platforms. AI added generated mockups and research summaries. Agents now draft specs, synthesize feedback, and build working prototypes, so the human work concentrates on judgment, taste, and deciding what not to build.
The shift: Agents draft specs, prototypes, and research synthesis; humans own taste and tradeoffs.
Copilot today
2.9 in five years
How has the product and design team changed across five eras?
Era 1 · On-prem
Before 2005
0.1Product managers wrote long requirement documents, designers worked in desktop apps and handed off static files, and research meant an occasional lab study with a handful of users. Engineering built what the document said, months later.
Feedback reached the team through sales anecdotes and the loudest customer, which made the roadmap a negotiation rather than an evidence base.
Era 2 · SaaS and cloud
2005 to 2020
0.7Cloud design tools, roadmapping software, product analytics, and A/B testing made the work collaborative and measurable. The product trio of PM, designer, and engineering lead became the standard team shape, with a researcher or data analyst shared across squads.
The roadmap became a living artifact, and so did the debate about it. A good example is the cybersecurity product roadmap, where the real work is choosing what to leave out.
Era 3 · AI-assisted
2020 to 2024
1.6AI generated mockups, copy, and first drafts of specs, and summarized interview transcripts and survey responses. Designers got faster at exploring options. PMs got faster at writing the document nobody wanted to read.
The team still decided what mattered and how it should feel. The tools sped up the middle and left both ends alone.
Era 4 · Agentic
2024 onward
2.0Agents now cluster feedback across tickets and calls, draft specs from it, and produce clickable or working prototypes that engineers can extend. A PM and a designer can test an idea in a day that used to take a sprint. Prototypes are cheap, so the constraint is choosing well; I argue that GEO itself is a product discipline for the same reason.
The new risk is volume. When anyone can ship a prototype, the team needs someone with the authority to say no.
Era 5 · Next 5 years
2026 to 2031
2.9My bet: product teams get smaller and more senior, with PMs and designers who prototype directly with agents and spend more time with customers. Taste, prioritization, and saying no become the core skills, and software gets generated for the moment rather than shipped as a fixed product.
Which product and design software is being rewired?
No deep dives yet. The first ones are in the queue below.
Coming next
- Product management
- Roadmapping
- User research
- Design tools
- Prototyping
- A/B testing
- Session replay
- Feedback management
Keep reading
Essays and analysis
What died (Tech Graveyard)
What comes next (Future Tech)
Comparisons
- Best AI Design Tools 2026: Canva vs Firefly vs Midjourney vs Recraft
- Top 5 Feature Flag Management Platforms of 2026: LaunchDarkly, GrowthBook, and More
- Top 5 AI Website Builders for 2026: Framer, 10Web, Durable, Wix, and Hostinger Compared
- Top 6 AI Coding Assistants of 2026: Cursor, Claude Code, Copilot, Codex, Windsurf and Tabnine Compared