AI writes the memo in nine seconds.
Deciding whether it's right is still your job.
Read the artifact below and make the call. Judgment Gym shows you what most people miss. Free public beta, no signup.
Approve the two weeks?
Traffic but no signups
The situation
A landing-page diagnosis and fix plan lands in your inbox, drafted by the growth team with an AI assistant from the analytics export. It asks for two weeks of design and front-end time.
Ines · Growth lead: “Design can start Monday if you sign off today. Anything you want to see first?”
You can pull one thing from the analytics export before you answer, then decide what to approve.
What you can see (4) · tap a card to read it
›E01 · Headline numbers, last 30 days
12,410 sessions, 11,890 unique visitors, 47 signups: a 0.38% conversion rate. Average session 31 seconds. Bounce rate 84%.
›E02 · The diagnosis
The document reads the short sessions and high bounce as visitors not finding what they need, cites industry benchmarks of 2–5% for SaaS landing pages, and concludes the page fails to communicate value quickly enough.
›E03 · The fix plan
Five changes: rewrite the hero headline (2 days), add a second and a sticky call to action (1 day), add customer logos and a testimonial (3 days), add a product demo video (1 week, needs design), cut the signup form from 5 fields to 2 (2 days). Expected result: 1.5–2% conversion, roughly 180–250 signups a month.
›E07 · What the export can do
The raw export has one row per session and per signup, with referral source, device and landing date. Ines says any split of signups, by source, device or day, would take her about an hour.
Step 1 of 2
What do you pull first?
Your first choice is kept. Changing your mind later counts as a retry.
01
Read the artifact
A launch memo, a roadmap tradeoff, an AI analyst's pricing recommendation. Written to look finished, with the problem buried where a confident reader skims past it.
02
Make the call
Pick an option and commit. No writing. The first choice is the one that counts; you can retry, and the retry is marked as one.
03
See what you missed
Authored feedback on the option you chose, the call the case recommends, and the evidence each rests on. Then the thing most people miss. That last part is the rep.
The library
18 cases, no signup
Critical Thinking & First Principles
The benchmark the model supplied: A plausible number, unsourced
Check one thing about the conversion assumption, then decide what goes to the board
Product Judgment
Two roadmap bets, one team: Which proposal gets the team?
Read one thing closely, then commit the team to A, B or a split
Working with AI: Delegation & Oversight
Automate it or keep it human: Switch the brief to the agent?
Ask one question, then adopt, modify or decline the agent workflow
One case a week, in your inbox
A new artifact to tear apart, plus what most people missed on the last one. No cohort, no curriculum — reps.
One email a week. Unsubscribe any time.