Insights
Research, practice and field notes on designing products that people and agents both have to get through.
For how we build Stunt Double, read Backstage.
Predictive UX Analysis for Product Teams: What It Is and When Simulation Beats Observation
Observational UX research tells you what already broke. Predictive UX analysis using agent simulation tells you what will break before it reaches real users. Here is how to know which one your team needs, and when.
· Michael Parker
What to Put in an AI Agent Testing Framework: The Decisions Before You Write a Single Script
Most teams building an AI agent testing framework start with tools. This piece argues the architectural decisions you make before writing a single script determine whether your framework catches real failures or just runs green.
Agent Identity Is Coming: How to Test Your Auth Flows Before the Standards Land
IETF and OpenID Foundation drafts on AI agent identity are already circulating, but no stable protocol is expected before December 2027. Here is how to test your auth flows against draft patterns now, before a forced migration makes it expensive.
Agent sessions look like bounces: how to tell them apart
· Michael Parker
The UX Metrics That Lie: Why Satisfaction Scores Miss Agent-Driven Experience Failures
CSAT and NPS can show 9/10 while AI agents silently fail to parse forms, loop on navigation states, and abandon tasks. Here is what those scores structurally cannot see.
Choosing an AI Agent Testing Approach: What UK Product Teams Actually Need to Decide
The choice between scripted automation, AI-assisted testing, and full agent simulation is not a tooling question. It is a question of what failure modes you are trying to catch, how fast your product changes, and who your users actually are.
Designing for AI Agents: What Product Teams Need to Know
AI agents are no longer a future concern — they're navigating your product today. Here's what product teams must understand about designing and testing interfaces built for both humans and agents.
· Michael Parker
Four Signals Your Product Analytics Will Never Surface (But Agent Simulation Will)
Analytics dashboards show where users dropped off. They cannot show where an AI agent mis-parsed a label, hesitated on a form, or abandoned a task without firing a single funnel event. These are the four blind spots.
MCP Is Already in Your Product: What Australian Growth Teams Need to Test Now
MCP-enabled agents are hitting Australian product flows right now, and most analytics stacks cannot tell the difference. Here is the audit checklist growth teams need before the gaps show in the data.
Why Your Regression Suite Misses Every Agent-Driven Failure Mode
Regression suites are built to catch what humans wrote breaking. They have no architecture for what an AI agent does next. Here is where that gap lives, and what a parallel test layer looks like.
Cursor AI: what it ships and what it misses
Cursor is the default AI-native IDE in 2026, shipping 41% of global code. But adoption and trust are moving in opposite directions. Here is what the numbers say.
What conversational AI actually does to your website
Conversational AI has moved past the chatbot. In 2026, agents act on your product without a user present, and most websites are not built for that. Here is what is changing, and what breaks first.