For a user who does not want to talk to a bot, it asks only for the email address and a brief issue description to create the support request.
What was measured
Response Completeness
Addresses all parts of the user’s support request instead of leaving key questions unanswered.
decisive for this rankingtransformation
A support bot that leaves key questions unanswered is not solving the customer’s issue. (3 of 3 judges)
What was given, what came back
Test input: Customer-support handoff request · text
Input — what we sent
The exact prompt
I received the wrong item in my order. I've already checked the order details and this is clearly a mistake on your end. I don't want any more back and forth — can you please connect me to a customer support agent or raise a ticket for this?
A frustrated support-escalation request after receiving the wrong item, asking to be connected to a support agent or have a ticket raised.
Why this input is hard
- · Human handoff reliability
- · Ticket creation workflow
- · Escalation contextual awareness
- · Professional tone under frustration
Output — unretouched
Provenance
- Observation
- ea008928-a23a-4a04-921c-e66a93a7729a
- Evidence run
- 2645dc92-49df-478a-b809-21dfd09f06a7
- Study
- Automate customer support using an AI chatbot
- Research task
- 86b9jm3ev
- Tested at
- not recorded
- Source
- first-party
- Evidence state
- verified
- Proof shown
- input + output shown
- Cost / latency
- not captured
- Repeat run
- not captured
- Tester
- not captured
The last three rows are honest blanks, not placeholders — our capture has no field for them yet.
Query this
get_evidence({
tool: "chatbase"
})MCP · mcp.aidemos.com/api/mcp
Free with attribution.
Same input, same check — 5 other tools
measured on Response Completeness
Freshdesk◐ MixedOn a double-charge fraud claim, the bot does not escalate immediately; it first frames the issue as a possible authorization hold and asks for transaction details, so the handoff is delayed rather than direct.FS Agent✓ WorkedCompletes a 'no bot' escalation by routing the user to support and requesting the email address and brief issue description needed to proceed.Respond✓ WorkedWhen the user said they did not want to talk to a bot, the bot handed the conversation to support and confirmed the connection.Wonderchat✓ WorkedIt fulfills a direct request for human help by listing multiple escalation channels — email, phone, and live chat — with response-time guidance instead of forcing self-service.Zendesk✓ WorkedThe bot escalates a fraud-like duplicate-charge complaint immediately and does not attempt to troubleshoot the payment itself.
This evidence is published in
Real inputs and real outputs, no retouching · every cell queryable via API & MCP · aidemos.com