Teardown: Grading Three Chatbot Opening Scripts
Objective: Apply the lesson's 5-step chatbot framework and its email-timing warning to grade three synthetic opening scripts, separating real defects from plausible-looking non-issues.
You're reviewing three draft chatbot scripts for Utkarsh Small Finance Bank's personal-loan landing page before launch: one written by a junior marketer, one by a sales-focused agency, one by the support team.
Grade each script's opener and question flow against the lesson's 5-step framework, flag real defects with severity, and don't flag things that only look like defects.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, structured way to grade multiple scripts consistently
Closest cataloged freemium chat tool since Drift and Intercom aren't individually in this site's directory
No access? Use Drift's or Intercom's own free trial (linked in this lesson's resources) to compare against a live playbook builder
The process
Specimens to review
Grade this script against the lesson's framework. What's the defect, and what specifically would you change?
SCRIPT A (junior marketer draft) Bot: "Hi there! Before we get started, what's your email address so I can send you our loan brochure?" Visitor: [drops off, no response] Bot: (no further messages, chat ends)
Specimen: synthetic, realistic
Grade this script. Is the generic opener the real defect, or is something else the bigger problem?
SCRIPT B (sales agency draft) Bot: "Hi! How can I help you today?" Visitor: "just looking at rates" Bot: "Great! What's your company size, role, and timeline for a decision?"
Specimen: synthetic, realistic
Grade this script. Does it have a real defect, or is this a clean pass?
SCRIPT C (support team draft) Bot: "Looking at personal loan rates? I can check your eligibility in under a minute, want to try?" Visitor: "sure" Bot: "Two quick things: roughly what loan amount, and is this urgent (next 30 days) or just exploring?" Visitor: "5 lakh, exploring for now" Bot: "Got it, since you're exploring, here's our EMI calculator and a quick eligibility guide, no forms needed. If you want a callback later just say the word."
Specimen: synthetic, realistic
Final deliverable
A graded teardown of all three scripts: defect (or 'clean pass'), severity, and the specific line to fix.
See a reference example
TEARDOWN LOG, Five-Star Business Finance MSME loan chatbot review Script A: CRITICAL, asks for phone number in message 1 before any qualifying value Script B: MODERATE, generic opener not specific to the top-up loan page it's deployed on Script C: CLEAN PASS, page-specific opener, 2 relevant questions, graceful disqualification with a resource offer
Success criteria
You're done when you can:
- Correctly identifies the email-timing defect in Script A as critical, citing the lesson's own warning
- Correctly identifies Script C as a clean pass without inventing a defect
- Does not flag any of the 4 distractors as defects