Chatbot Lab
Chatbot Lab

YOUR WORKING TOOL

Build a scorecard

Set your own priorities. Score observed behaviour, not marketing claims.

TRY A WORKED SCENARIO

Start with an example

Fictional inputs, explained below. Loading an example replaces the form. Your edits stay in this tab.

01

Promising demo, incomplete evidence

Two dimensions observed; four remain unknown. The result must stay provisional.

02

High average, failed pause

Strong scores cannot cancel a failed mandatory control. See the blocker alongside the average.

03

A documented first trial

All dimensions and mandatory checks entered. These fictional scores illustrate the calculation only.

Your observations

Adjust the inputs to match the workflow you are planning.

Leave a dimension untested until you have evidence. Score observed behaviour from 0 (failed) to 5 (consistently met your criterion); weights of 0 exclude a dimension.

Context
Voice consistency
Controls
Handover
Billing clarity
Data handling
Mandatory trial checks

An average cannot override a failed check. These are your observations, not automatic verification.

WORKING RESULT

Your scenario

Adjust the inputs to see your result. The method and assumptions are described below.

No account connection. No messages sent. Your inputs are processed in this tab.

TESTS & EVIDENCE

Lab notes

Better questions. Repeatable tests. Evidence you can keep beside the score.

Explore the blog