Does Dot keep the next question tied to the previous answer?
Dot kept the follow-up tied to the five customers in its previous answer. It reported 86 recent orders for that group, without asking who they were or recalculating the spend ranking.
1 of 1 test case passed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
"They" refers to the previous answer's setPassEvidence
Input 1 · Initial ranking question: “Who are our top 5 customers by total spend?”
Input 2 · Follow-up order question: “How many orders did they place in the last 90 days?”
Follow-up answer They placed 86 orders combined in the last 90 complete days, July 19–October 16 . Order coverage note This counts all orders placed, including those later cancelled or refunded, unlike the spend ranking. Initial customer table name customer_id currency total_spend_usd order_count earliest_order_date latest_order_date 1 Dev Raghavan 17 USD $20,222 19 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 2 Priya Raghavan 6 USD $17,203 17 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 3 Rosa Delgado 11 USD $13,904 17 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 4 Jae-won Park 10 USD $13,647 18 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 5 Cassandra Pope 15 USD $13,107 18 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 Follow-up customer table name order_count 1 Dev Raghavan 18 2 Priya Raghavan 17 3 Rosa Delgado 19 4 Jae-won Park 14 5 Cassandra Pope 18
Transcribed from Proof 1 · Screen recording at 0:21
Transcribed from Proof 1 · Screen recording at 0:41
Transcribed from Proof 1 · Screen recording at 0:42
Everything the tool returned, in order (6)
- Initial answer
Dev Raghavan leads with $20,222 in total spend . The chart shows your top five customers.
- Follow-up answer
They placed 86 orders combined in the last 90 complete days, July 19–October 16 .
- Order coverage note
This counts all orders placed, including those later cancelled or refunded, unlike the spend ranking.
- Query behind the follow-up answer · not part of the answer
cte_order_counts AS ( SELECT customer_id, COUNT(id) AS order_count FROM customer_support_db.public.orders WHERE customer_id IN (17, 6, 11, 10, 15) AND (placed_at AT TIME ZONE 'America/New_York')::date >= '2026-07-19'::date AND (placed_at AT TIME ZONE 'America/New_York')::date <= '2026-10-16'::date GROUP BY customer_id ) SELECT tc.name, COALESCE(oc.order_count, 0) AS order_count FROM cte_target_customers tc LEFT JOIN cte_order_counts oc ON tc.customer_id = oc.customer_id ORDER BY tc.rank_order ASC
- Initial customer table
name customer_id currency total_spend_usd order_count earliest_order_date latest_order_date 1 Dev Raghavan 17 USD $20,222 19 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 2 Priya Raghavan 6 USD $17,203 17 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 3 Rosa Delgado 11 USD $13,904 17 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 4 Jae-won Park 10 USD $13,647 18 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519 5 Cassandra Pope 15 USD $13,107 18 2025-10-15 22:17:00+00:00 2026-10-12 21:08:52.519
- Follow-up customer table
name order_count 1 Dev Raghavan 18 2 Priya Raghavan 17 3 Rosa Delgado 19 4 Jae-won Park 14 5 Cassandra Pope 18
| ✓ | Expected: Answer for the previous reply's customer group, without asking who they are or silently changing its membership.Found: The second table kept the original five customers in order and returned 18, 17, 19, 14 and 18 orders respectively. |
The follow-up needed to preserve the original customer group. Dot's query reused those five customers in their original rank order rather than selecting a new group, keeping the reference intact.
Different order coverage. The spend ranking excluded cancelled and refunded orders; the follow-up counted all orders placed, including those later cancelled or refunded.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | AI Database Agents → | 28 scenarios · 11 tools |
| Capability | Conversational Interaction → | |
| Scenario | The next question refers to the previous answer → | |
| Tool | Dot → |
Global scenario definition → · Global capability definition → · Dot product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.