Can Basedash understand follow-up questions that refer to the previous answer?
Basedash kept the previous answer’s set in a follow-up question. It counted 86 orders for the same five customers without asking who they were.
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: “Who are our top 5 customers by total spend?”
Input 2: “How many orders did they place in the last 90 days?”
Response 1 answers Input 1.
Response 2, the final response, answers Input 2.
Those five customers placed 86 orders combined in the last 90 days. Customer | Orders; Dev Raghavan 18; Priya Raghavan 17; Rosa Delgado 19; Jae-won Park 14; Cassandra Pope 18; Total 86. Period: July 15–October 12, 2026 (UTC). These counts include cancelled orders, since they were still placed; the earlier spend ranking excluded cancelled orders.
Transcribed from Proof 1 · Screen recording at 0:11
Every message the tool sent, in order (2)
- Top five customers by spend
Your top 5 customers by total spend are: Rank | Customer | Total spend | Orders; 1 Dev Raghavan $21,164.27 20; 2 Priya Raghavan $17,202.51 17; 3 Rosa Delgado $14,659.75 18; 4 Jae-won Park $13,646.96 18; 5 Cassandra Pope $13,106.87 18. Total spend is the sum of all-time order totals, excluding cancelled orders. It does not subtract refunds or include subscription or membership fees.
- Orders placed by the five customers
The final response, printed above.
| ✓ | Expected: Answer the follow-up for the same set as the first answer, without asking who they are or changing the set.Found: The follow-up counted 86 orders for the same five customers, in the same order, without asking who they were. |
The follow-up needed to use the first answer’s set without asking again or changing it. Basedash answered for the same five customers in the same order and asked no clarifying question.
Different counting rules. The follow-up count included cancelled orders; the earlier spend ranking excluded them.
Set carried into query. The follow-up’s generated query also carried over the customer set, not just the answer text.
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 | Basedash → |
Global scenario definition → · Global capability definition → · Basedash 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.