Skip to main content
Build a Conversation AI bot that engages new leads the moment they come in, asks qualifying questions, scores them based on their answers, tags them in your CRM, and routes hot prospects directly to your sales team — all within minutes of first contact.
Overview

Bot goals setup screen showing conversation objectives and qualification targets

Bot goals configuration for defining qualification objectives

Architecture overview

Here is how the lead qualification flow works:
  1. Lead enters the system — A new contact arrives via form submission, web chat, SMS opt-in, Facebook ad, or manual import.
  2. Bot initiates conversation — A workflow triggers the Conversation AI bot to send the first message within seconds.
  3. Qualifying questions — The bot asks a series of questions to determine fit: budget, timeline, needs, decision-making authority.
  4. Lead scoring — Based on the answers, the bot assigns a score using custom field updates and tags.
  5. CRM update — The bot writes answers to custom fields on the contact record and moves the opportunity to the appropriate pipeline stage.
  6. Routing — Hot leads get an immediate notification to a sales rep. Warm leads enter a nurture sequence. Cold leads are tagged for future marketing.

Step-by-step build

1

Create custom fields for qualification data

Navigate to Settings > Custom Fields and create the following fields on the Contact object:
Use dropdown fields wherever possible instead of free text. This makes filtering, reporting, and workflow conditions much easier.
2

Set up your CRM pipeline

Go to Opportunities > Pipelines and create a pipeline (or modify an existing one) with these stages:
  1. New lead — Contact just arrived, not yet engaged
  2. Qualification in progress — Bot is actively qualifying
  3. Qualified - Hot — High score, ready for sales
  4. Qualified - Warm — Medium score, needs nurture
  5. Qualified - Cold — Low score, long timeline or low budget
  6. Disqualified — Not a fit
  7. Contacted by sales — Sales rep has reached out
For pipeline setup details, see pipeline opportunities.
3

Create the Conversation AI bot

Navigate to AI Agents > Conversation AI and click Create Bot. Configure:
  • Name: “Lead Qualifier”
  • Type: Suggestive (recommended for initial setup) or Auto-pilot once tested
  • Channels: Toggle on the channels where leads arrive — typically SMS, Web Chat, and Facebook Messenger
  • Calendar: Leave unassigned for now (this bot qualifies, not books)
See Conversation AI setup for detailed bot creation steps.
4

Write the bot prompt

Paste the full prompt from the section below into the System Prompt field. The prompt instructs the bot to:
  • Greet the lead and set expectations
  • Ask qualifying questions one at a time (never dump all questions at once)
  • Listen for buying signals
  • Summarize what it learned
  • Hand off to the appropriate next step
The bot must ask questions conversationally, not like a survey. Leads who feel interrogated will disengage. The prompt includes specific instructions for natural transitions between questions.
5

Configure bot actions for CRM updates

In the bot’s Actions settings, configure Conversation AI actions to update the contact record based on conversation content:
  • Update custom field: Map each qualifying answer to its corresponding custom field
  • Add tag: Add tags like “qualified-hot,” “qualified-warm,” or “qualified-cold” based on the conversation outcome
  • Move opportunity: Advance the opportunity to the appropriate pipeline stage
  • Assign user: Route hot leads to a specific sales rep or round-robin assignment
Appointment action setup showing calendar selection and booking options

Appointment booking action configuration for scheduling qualified leads

6

Build the lead scoring workflow

Go to Automation > Workflows and create a workflow that calculates the lead score after qualification is complete.Trigger: Tag added — “qualification-complete”Scoring logic (use If/Else branches):After scoring:
  • Score 70–100: Tag “qualified-hot,” move to “Qualified - Hot” stage, send internal notification
  • Score 40–69: Tag “qualified-warm,” move to “Qualified - Warm” stage, enroll in nurture sequence
  • Score 0–39: Tag “qualified-cold,” move to “Qualified - Cold” stage
See customizing workflow actions for step-by-step workflow building.
7

Create the hot lead alert

Add a final step to the scoring workflow for hot leads:
  1. Internal notification — Send an email and/or SMS to the assigned sales rep with the lead’s name, score, and qualification summary
  2. Task creation — Create a task for the sales rep to call the lead within 15 minutes
  3. Bot handoff — Stop the Conversation AI bot on this contact so the sales rep can take over manually
Speed to contact is critical. Studies show that leads contacted within 5 minutes are 21x more likely to convert. Your hot lead alert should trigger immediate action.
8

Build the nurture sequence for warm leads

Create a separate workflow for contacts tagged “qualified-warm”:
  1. Day 0: Send a value-add email (case study, guide, or resource relevant to their pain point)
  2. Day 3: Follow-up SMS checking if they have questions
  3. Day 7: Send a second resource or testimonial
  4. Day 14: Re-engagement message with a soft call to action
  5. Day 21: Final outreach before moving to cold
If the lead re-engages at any point, the bot can re-qualify and update the score.

Full prompt


Testing your qualification bot

1

Hot lead simulation

Submit a test lead and respond with high-budget, immediate-timeline, decision-maker answers. Verify the bot scores them as hot, the internal alert fires, and the opportunity moves to the correct pipeline stage.
2

Warm lead simulation

Respond with medium-budget, 1–3 month timeline answers. Verify the bot tags and routes correctly and the nurture sequence enrolls.
3

Cold lead simulation

Respond with low-budget, 6+ month timeline, not the decision maker. Verify the contact is tagged cold and no sales alert fires.
4

Resistance test

Refuse to answer the budget question. Verify the bot handles it gracefully and does not loop or break.
5

Human handoff test

Say “Can I talk to a real person?” Verify the bot immediately stops and routes to a team member.
6

Multi-channel test

Run the qualification flow on SMS, web chat, and Facebook Messenger. Verify the bot adjusts its message length and tone for each channel. See multi-channel deployment for channel-specific guidance.
7

CRM verification

After each test, check the contact record. Verify all custom fields are populated, tags are applied, the opportunity is in the correct stage, and the lead score is accurate.

Optimization tips

  1. Analyze drop-off points. Review conversation transcripts to find where leads stop responding. If most drop off at the budget question, soften the framing in your prompt.
  2. A/B test opening messages. Try different first messages — question-based versus statement-based — and track which gets higher response rates.
  3. Refine scoring weights. After 30 days, compare your lead scores against actual close rates. Adjust the scoring weights so the model better predicts which leads actually buy.
  4. Shorten the flow for repeat visitors. If a contact has already been qualified, update the prompt to skip previously answered questions and focus on what has changed.
  5. Add industry-specific questions. Generic qualification works, but adding one or two questions specific to your industry dramatically improves lead quality.
  6. Monitor response time. The bot should respond to the first message within 60 seconds. If there are delays, check your workflow trigger configuration.

KPIs to measure success


Next steps

Appointment scheduler

Add appointment booking to your qualification flow so hot leads can book immediately.

Conversation AI actions

Deep dive into all available bot actions for CRM updates.

Workflow automation

Build more sophisticated scoring and routing workflows.

Prompt engineering

Advanced prompt techniques for better qualification conversations.
Last modified on March 5, 2026