The review for this task
Can you send an AI-built weekly report to your boss?
Growth reports · five capabilities, three tools
ChatGPT scored 95.5 overall and Quadratic 74.8; both got every company metric right, and ChatGPT led on the questions
Each row is one part of a capability. Filled squares are right answers, outlined squares wrong ones. The overall score is the average of the five capabilities.
RightReported found by the tool; no rows to checkPartly rightWrongNot tested or not visible in the record
| Capability | ChatGPT | Quadratic | Julius |
|---|---|---|---|
| 1 · Metrics with stated rulesCompany metrics, first round; one square per metric value | 72/72 | 72/72 | 69/693 not visible in the record |
| 1 · Metrics with stated rulesBasic questions (1 to 10); one square per question | CSV run · 10/10Excel run · 10/10 Best | CSV run · 9/10Excel run · 9/10 | 9/10 |
| 2 · Data refresh and recalculationCompany update and recalculation; one square per company | 6/6 | 6/6 | 4/51 partly right · 1 not tested |
| 3 · Cross-table business rulesCross-table questions (11 to 40); one square per question | 27/30Best | CSV run · 21/30Excel run · 11/30 | Not run (out of credits) |
| 4 · Data quality checksEvent check (48 planted errors); one square per planted error | 48/48 | 48/48reported; 8 rows checked | 40/408 not visible in the record |
| 4 · Data quality checksDirty-data questions (41 to 50 and 58); one square per question | 11/11Best | CSV run · 8/11Excel run · 6/11 | Not run (out of credits) |
| 5 · Business interpretationBusiness-judgment questions (51 to 57 and 59); one square per question | 7/8Best | CSV run · 5/8Excel run · 2/8 | Not run (out of credits) |
- Julius has no overall score: it ran out of credits before question 11, so three capabilities were not run.
- Partly right: in one company update Julius had the right numbers but called a changed refund metric unchanged.
- Red marks the best result in a row only where the tools can be compared; ties are marked equally.
3 products · 6 fictional companies + a 60-question bank · all data synthetic · tested 2026-09-23 to 2026-10-12analyst.ai
Not without checking it. ChatGPT scored 95.5 overall and Quadratic 74.8. Both got every company metric right; ChatGPT led on questions where the AI had to set the rules. Julius ran out of credits after 10 questions.