Website chatbot lead qualification checklist

Our test is simple: the visitor should understand why the next question is being asked. If we cannot explain that, we remove it.

Use this checklist before launch and after every material change.

Define the job

Write one sentence: “The chatbot helps this visitor reach this next step.” Examples include booking a property consultation, requesting a manufacturing quote, or scheduling an attorney intake call.

Choose one primary conversion. Multiple competing paths make the conversation longer and the reporting less useful.

Approve the knowledge

List the exact pages, documents, and records the bot may use. Remove expired prices and conflicting policies. Assign a person to approve changes.

Mark restricted subjects. The bot should say when it does not know and offer a person instead of improvising.

Ask only qualification questions that change the route

Every question should affect eligibility, destination, urgency, or preparation for the next conversation. If the answer changes nothing, remove the question.

Collect contact details after the visitor receives some value. Explain why each sensitive field is needed.

Build the handoff

Define triggers for human takeover: explicit request, complaint, uncertainty, high-value opportunity, restricted topic, repeated misunderstanding, or system error.

The employee should receive the transcript, captured details, source page, and reason for handoff. Do not force the visitor to repeat the conversation.

Connect the calendar carefully

Offer only the calendar that matches the qualification path. Confirm timezone, duration, location, and rescheduling. Test daylight-saving changes and full calendars.

A booking is not finished until the CRM records it and the visitor receives accurate confirmation.

Disclose the interaction

Identify the chatbot as automated or AI at the start. State how submitted information will be used and link to the privacy notice. Do not imitate a named employee.

For legal and other sensitive businesses, disclosure and data handling need industry review.

Test failure cases

Submit blank fields, false names, long messages, another language, a request outside service area, a price exception, profanity, and a prompt asking the bot to ignore its rules. Verify refusal, recovery, and handoff.

Repeat a previous visitor and confirm that duplicates are handled without exposing another person’s information.

Measure useful outcomes

Track answered conversations, qualified conversations, bookings, handoffs, unanswered questions, incorrect answers, opt-outs, and completed sales. Read a sample of conversations every week.

Do not celebrate total chats without knowing whether the visitors reached the right next step.

Assign maintenance

Name the owner for knowledge updates and production incidents. Set a response time. Record changes to prompts, data sources, integrations, and models.

DATIK’s Lead Chatbot includes setup and monthly operation. The published price and included period are on the pricing page.

Sources and review limits

• NIST Generative AI Risk Management Profile
• U.S. SBA guidance on AI for small business

Sources were checked on September 25, 2026. This checklist does not replace privacy, security, accessibility, or industry-specific legal review.