The problem with unsorted lead intake
When every web form submission lands in the same CRM queue, sales teams face three recurring issues:
- Clogged schedules: Agents spend prime calling hours on contacts who are months away from making a decision.
- Buried high-intent buyers: Motivated buyers with short timelines get lost in a crowded, unsorted inbox.
- Guesswork on follow-up: Without timeline or intent data, reps have no clear signal for who to call first.
The fix is not more manual sorting — it is an automated qualification pipeline that reads lead intent on arrival and routes contacts accordingly.
How a context-aware qualification pipeline works
A context-aware pipeline connects your lead intake form to an AI reasoning layer and a structured CRM. Here is how the three pieces fit together:
1. Intake form captures intent signals
Add two or three fields to your existing form: estimated move-in or purchase timeline, budget range, and property type. These fields feed the pipeline without requiring a phone call.
2. OpenAI scores the lead
When a new submission arrives, a webhook triggers an OpenAI API call. The prompt passes the lead's raw form data and asks the model to return a structured JSON object with three fields:
readiness_tier: immediate (0–30 days), near-term (31–90 days), or long-term (90+ days)priority_score: a 1–10 integer based on stated timeline and budget specificityrouting_action: escalate, nurture, or stage
A system prompt defines the scoring criteria so the output is consistent across all submissions.
3. Airtable receives and routes the result
The pipeline writes the scored lead to an Airtable base. Airtable automations then branch on readiness_tier:
- Immediate: Triggers a Slack or SMS notification to the assigned agent with the lead's contact details and form responses.
- Near-term: Enrolls the contact in a three-touch email sequence spaced over two weeks.
- Long-term: Adds the contact to a low-frequency drip with a 60-day check-in task.
Building the pipeline step by step
Step 1: Set up the Airtable base
Create a table called Leads with fields matching your form outputs plus the three AI-scored fields: readiness_tier, priority_score, and routing_action. Add a Status single-select field with options: New, Escalated, Nurture, Staged.
Step 2: Write the OpenAI prompt
Keep the prompt brief and deterministic. Example:
You are a lead qualification assistant for a real estate team.
Given the following lead form submission, return a JSON object with:
- readiness_tier: "immediate", "near_term", or "long_term"
- priority_score: integer 1-10
- routing_action: "escalate", "nurture", or "stage"
Base readiness_tier on the stated move-in timeline.
Base priority_score on how specific the budget and timeline are.
Request response_format: { type: "json_object" } in your API call to guarantee parseable output.
Step 3: Connect the webhook
Use Make, n8n, or a lightweight serverless function to:
- Receive the form POST
- Call the OpenAI Chat Completions endpoint with the form data
- Parse the JSON response
- Write a new record to Airtable with all form fields plus the three scored fields
Step 4: Configure Airtable automations
Build one automation per routing action. Each automation triggers on Status = New and checks routing_action:
- If
escalate: send a notification and set Status to Escalated - If
nurture: add to email sequence and set Status to Nurture - If
stage: create a follow-up task 60 days out and set Status to Staged
What this changes operationally
Once the pipeline is live, agents receive only pre-screened, escalated leads in their real-time notifications. Long-term contacts move through automated sequences without manual intervention. The CRM stays clean because every record carries a machine-assigned tier from the moment it enters the system.
The pipeline does not replace agent judgment — it removes the work of sorting so agents can focus judgment where it matters: on conversations with buyers who are ready to move.
Where to go from here
Start with a single lead source and one form. Get the OpenAI scoring working and verify the JSON output is consistent across 20 to 30 test submissions before connecting the full Airtable automation. Accuracy on the readiness_tier field is what determines whether the routing logic holds up at scale.
Once the single-source pipeline is stable, the same pattern applies to any other intake channel — referrals, open house sign-ins, or inbound calls transcribed by a voice AI.