- Launch Studio
- Profound
- Screaming Frog
- Lighthouse
- Sheet
Case study · Launch Studio on AEM Edge Delivery
One SEO and GEO compliance finding on a pharma site for healthcare professionals, taken from scan to live in an afternoon. Then the same fix priced through a conventional stack.
A branded HCP site on AEM Edge Delivery Services.
The Optimize module in Launch Studio crawled the page and found four SEO findings and eleven GEO findings.
One click, Send to Plans, turned the fifteen findings into nineteen tasks with a definition of done on each. Nobody wrote a ticket.
How to read the comparison. Each stage below shows Launch Studio on top, as recorded. The conventional stack sits in a drawer under it: a hover or tap pulls out the people, tools, hands-on time and cost the same stage takes with Profound, a crawler audit, GitHub, a developer driving Claude Code and Edge Delivery.
The numbers, stage by stage
Launch Studio on top. Hover or tap a stage to pull out what the conventional stack takes.
Estimates for the conventional column, issue by issue. Six of the fifteen are code changes; nine are content changes. Client approval and regulatory review are left out of both columns, since they are the same step whichever way the work is done. Hands-on hours include writing or development, review, QA and the producer's coordination. Where Flight gave a number in the session, the Launch Studio column uses it: the three remaining Content Structure tasks were estimated at 15 to 20, 30 to 40 and 20 to 30 minutes, plus 10 to 15 for the group's QA, 75 to 105 minutes in all with nobody waiting on the user.
Missing <article> elementDone in the session
Heading hierarchy skips levels
WebSite schema missing potentialAction
Article, Person and Organization schema (recommendation #1)
Two further findings not visible in the recorded screens (Ongoing)
No authoritative citations, statistics without sources, unsourced stats under E-E-A-T (three findings, one job)
Low answer-first architecture (14% of sections open with a direct answer)
No author byline
Low citation diversity (3 citations, 2 domains)
No answer islands (no 134 to 167 word self-contained blocks)
No first-person experience signals
Stale year references
One further Medium finding not visible in the recorded screens
Conventional totals for all fifteen
Twenty-eight screens from the session behind these numbers: the scan, the plan, the Dev Agent at work, and the ten moments a person stepped in. Register with your work email to open it.