myAI Scouts

~8 min

Multi-team experiment requirements — before you build something that touches everyone's turf

Sent to senior ICs and team leads — the day-to-day owners who make implementation decisions — across product, UX design, front-end engineering, data analytics, brand/marketing, legal/compliance, and customer success at a B2B SaaS or e-commerce company. These participants own specific parts of a checkout flow, onboarding sequence, or pricing page that a proposed experiment would touch. They have authority to raise a block or flag a dependency before the experiment is scoped in Confluence or ticketed in Jira. Findings are used to catch every team's hard constraints before a single Jira ticket is created — preventing the blocked sprints and last-minute legal holds that happen when requirements surface mid-build.

What this scout explores

The constraint from a past experiment they wish had been caught earlier

Think about a previous experiment or product change that affected your team's area — a checkout redesign, a copy test, a pricing page change. Was there a constraint, compliance requirement, or tracking dependency your team had that wasn't captured before the build started? What happened as a result? This is the most important thing to surface before scoping.

Their hard constraint on this specific experiment

What must this experiment not do within the area your team owns — a brand guideline it must respect, a compliance rule it cannot breach, an Amplitude event it cannot break, a customer promise it must uphold? What would cause your team to raise a Jira block if it wasn't addressed in the design?

The dependency they're assuming has been confirmed

What does your team need from another team for this experiment to work — a data schema change from engineering, a legal sign-off from compliance, a new Amplitude tracking event, updated copy from brand? Is that handoff confirmed in writing, or is it still assumed?

What would make them raise a hold the week before launch

Under what specific circumstance — a missing legal review, a broken analytics event in staging, a brand inconsistency that gets past QA — would your team escalate a hold the week before this experiment is scheduled to go live?

This scout will be copied to your account as a draft. You can customise it before activating.