Lead qualification workflow
- LEAD FORM
- WEBHOOK
- NORMALIZE
- LOCAL QWEN ANALYSIS
- STRUCTURED RESULT
- DETERMINISTIC SCORE
- HOT / WARM / COLD
- POSTGRESQL
- TELEGRAM / EMAIL
AUTOMATION USE CASE / 01
AI LEAD QUALIFICATION & ROUTING
LOCAL AI / DETERMINISTIC SCORING / AUTOMATED ROUTING
A workflow that analyses incoming leads with a local language model, converts qualitative signals into deterministic scoring and automatically routes leads based on priority.
- STATUS
- WORKING DEMO
- TYPE
- SALES AUTOMATION
- TECH
01 / PROBLEM
EVERY LEAD SHOULD NOT REQUIRE THE SAME MANUAL REVIEW
When every enquiry lands in the same inbox, someone still has to read it, evaluate intent, estimate urgency and decide what should happen next.
The workflow was built to demonstrate how this first qualification step can be structured and automated without handing the entire decision to an opaque AI model.
02 / SOLUTION
AI ANALYSIS + EXPLICIT BUSINESS RULES
The language model analyses qualitative signals in the incoming enquiry.
The final numerical score and priority are calculated by deterministic rules inside the workflow.
AI ANALYSES.RULES SCORE.WORKFLOW ROUTES.
03 / WORKFLOW
FROM ENQUIRY TO ACTION
04 / AI ANALYSIS
QUALITATIVE SIGNAL EXTRACTION
01
PURCHASE INTENT
Signals that the enquiry is about purchasing a service.
02
REQUIREMENTS CLARITY
How clearly the enquiry describes the requested work.
03
PROJECT FIT
How well the requested project matches the service.
04
URGENCY
Signals about the timeline and need for a response.
The model does not return the final business priority as an unquestioned decision.
It produces structured qualitative information which is passed into deterministic scoring rules.
05 / DETERMINISTIC SCORING
INSPECTABLE SCORING RULES
The score from 0 to 100 is calculated using explicit rules. The scoring categories include purchase intent, budget, requirements clarity, project fit and urgency.
- 80–100
- HOT
- 40–79
- WARM
- 0–39
- COLD
06 / EXAMPLE RESULTS
THREE DEMO ENQUIRIES
HOT
100 / 100
High intent, clear requirements, suitable budget and urgent timeline.
WARM
45 / 100
Potential opportunity, but requirements or buying intent remain incomplete.
COLD
18 / 100
Low purchasing signal or insufficient project information.
These are demo examples, not fixed labels.
The HOT / WARM / COLD preset buttons used in the demo only populate sample form data. They do not directly control the classification result.
07 / VIDEO DEMO
SEE THE WORKFLOW IN ACTION
VIDEO DEMO
MEDIA PENDING
08 / ROUTING & ACTIONS
PRIORITY-BASED FOLLOW-UP
HOT
IMMEDIATE NOTIFICATION
Immediate notification / priority follow-up.
WARM
NORMAL FOLLOW-UP
Normal follow-up / review queue.
COLD
LOWER-PRIORITY PROCESSING
Lower-priority processing.
The current demo stores leads in PostgreSQL and delivers notifications through Telegram and Gmail.
09 / CUSTOMIZATION
THE ROUTING LAYER CAN BE ADAPTED
Can be connected to a CRM, Slack, Microsoft Teams, SMS, Google Sheets, HubSpot or custom APIs. These are possible extensions, not integrations currently implemented in this demo.
10 / BUSINESS USE