Every new lead from a form is automatically scored on budget, need and timing and reaches the right salesperson within minutes, complete with a priority score and rationale.
automatinganalytical
Description
The Lead Relay takes over the first review of new leads as soon as they come in through a form. Instead of every inquiry being screened manually and prioritized roughly by gut feeling, an AI scoring step evaluates budget, need and timing against predefined criteria and assigns a traceable A to D score with a rationale.
The core of the workflow is the point logic itself: three criteria, each worth up to ten points, and an overall score that determines how the lead is handled next. What sets it apart from a plain scoring tool is the confidence rating. When form fields are incomplete or an entry stays vague, the assistant flags this in the result instead of projecting a precision the data does not support. That builds trust in the prioritization, especially at the borderline between two scoring tiers.
A human stays in the loop at two points: calibrating the scoring thresholds initially against real deals already won or lost, and spot checking the first runs before the workflow continues without supervision. From there, the relay handles the presorting and sales handles the close.
Fast response: Promising leads reach the right salesperson within minutes, instead of sitting in an inbox for days.
Consistent prioritization: Every lead is scored against the same criteria, no matter who sees the inquiry first.
Focus on the promising cases: A and B leads are immediately visible, while C and D leads flow into nurturing.
Clean CRM data: Every lead is filed with a score, rationale and status, not just the raw data from the form.
Steps
Every step shows who carries it out: icon, colour and label together indicate whether a person acts, whether it runs automatically, whether a result is produced, or whether an approval is required.
01 · Person
A form webhook delivers the incoming lead data, including budget, need and timing, to the workflow.
02 · Automated
Company size, industry and tech stack can optionally be added through an enrichment service such as Clearbit or Apollo, to complete the picture before scoring.
03 · Automated
Budget, need and timing are each scored from zero to ten points, resulting in an overall score from A to D with a rationale.
04 · Automated
The workflow checks whether the lead already exists in the CRM and either creates a new entry or updates the existing one with the score and rationale.
05 · Result
The responsible person receives the score, rationale and a concrete recommended action directly in team chat, for example via Slack, Teams or email.
Key
PersonAutomatedResultApproval
Requirements and operations
Requirements
Form toolA tool with a webhook trigger, for example Typeform or Google Forms.
CRM systemHubSpot, Salesforce, Pipedrive or a comparable system for storing leads.
Scoring criteriaThresholds for budget, product fit and timing, defined before go live.
LLM access and team chatAccess to a language model, for example from OpenAI, Anthropic or Google, plus a chat tool such as Slack or Teams for the sales notification.
Testing
Test leadSubmit a form with realistic data, including budget, need and timing.
Data mappingCheck whether all form fields arrive completely in the workflow.
Scoring and CRMCheck whether the score is plausible and the lead lands with the right status in the right stage.
Edge casesTest incomplete data, a duplicate case, a very small budget and recognizable spam.
Go-live
ApproveActivate the workflow only once scoring, CRM entry and routing run cleanly.
Cross checkManually review the first roughly twenty scored leads and sharpen the thresholds.
RoutingRoute leads by region, industry or budget to the right salesperson.
ExpandAdd enrichment, nurturing tracks for B and C leads, and intent analysis as needed.
Common issues
Form fields not mappedConnect every placeholder in the prompt explicitly to the matching form field name.
No valid JSONKeep the instruction for pure JSON output in the prompt and add validation.
CRM required fields missingSet email address, last name and other required fields in the form as mandatory.
Scoring feels arbitrarySpell out the thresholds explicitly in the prompt and calibrate them against real deals.
Builder Prompt
Copy the prompt below in full into your automation tool or agent builder. As a file: lead-staffel.en.json
GOAL
Build a workflow that automatically scores new leads from forms on budget, need and timing, assigns an A to D score with a rationale, files the lead cleanly in the CRM, and notifies the responsible salesperson with a concrete recommended action.
Adapt the structure to the concepts of your automation tool, for example nodes, steps, zaps or agent actions. While doing so, preserve the trigger logic, the sequence, the parallelism and the data flow.
TRIGGER
Type: form webhook, for example Typeform or Google Forms. Fires on every new submission.
Requirement: required fields in the form for email, name, budget, need and timing, otherwise the CRM is missing required data.
STEPS
1. Optional, enrichment: add company size, industry and tech stack through Clearbit, Apollo or a comparable service.
2. AI step: lead scoring. Use exactly the scoring prompt below. The output is a JSON with total_score, scoring_category, the three sub scores, qualification_status, recommended_action and summary.
3. Duplicate check: search for the lead by email address or company in the CRM. Update an existing entry with the new score, instead of creating a duplicate.
4. CRM: create or update the lead, with score, category, rationale and status in the right stage.
5. Output: notify the responsible salesperson, with routing by region, industry or budget. The message contains the score, rationale and recommended action. A and B leads are reported immediately and prominently, C and D leads flow into nurturing.
PROMPTS (use exactly and unchanged)
Prompt 1, lead scoring:
"""
You are an assistant for lead qualification. Analyze the following lead data and assess the lead quality:
Lead information:
- Name: [LEAD_NAME]
- Company: [COMPANY_NAME]
- Budget: [BUDGET_INFO]
- Need/requirement: [NEED_DESCRIPTION]
- Timing/timeframe: [TIMING_INFO]
- Additional information: [ADDITIONAL_INFO]
Your task:
Assess the lead against the following criteria and assign an overall score.
1. Budget fit (0 to 10 points): [BUDGET_CRITERIA]
2. Need fit (0 to 10 points): [NEED_CRITERIA]
3. Timing fit (0 to 10 points): [TIMING_CRITERIA]
Scoring scheme:
- A lead (25 to 30 points): high priority, immediate contact
- B lead (18 to 24 points): good quality, contact within 24 hours
- C lead (10 to 17 points): medium quality, nurturing required
- D lead (0 to 9 points): low quality, automatic follow-up
Respond exclusively in the following JSON format:
{
"total_score": [total points],
"scoring_category": "[A/B/C/D]",
"budget_points": [0 to 10],
"need_points": [0 to 10],
"timing_points": [0 to 10],
"qualification_status": "[Qualified/Conditionally qualified/Not qualified]",
"recommended_action": "[concrete recommended action]",
"summary": "[2 to 3 sentences justifying the score]"
}
"""
DATA FLOW
Trigger/step 1 to step 2: form fields to [LEAD_NAME], [COMPANY_NAME], [BUDGET_INFO], [NEED_DESCRIPTION], [TIMING_INFO]. Enrichment data to [ADDITIONAL_INFO].
Step 2 to step 4: JSON fields to the CRM properties (score, category, status, rationale).
Step 2 to step 5: scoring_category for routing and prioritization, recommended_action and summary.
ERROR HANDLING
Form fields not mapped: connect every placeholder explicitly to the form field name, instead of guessing.
AI returns no valid JSON: keep the pure JSON output instruction, validate the output and repeat the step once. After that, file the lead unscored in the CRM and send a note to sales.
Incomplete lead data: have missing information scored as "not provided", instead of guessing it.
Spam or fake entries: treat these as a D lead, or filter them out through a plausibility check on the email domain.
Duplicate found: update the existing entry, instead of creating a new one (step 3).
Scoring feels arbitrary: fill in the three criteria placeholders concretely, calibrate them against real deals, and manually cross check the first twenty leads.
PLACEHOLDER CONVENTION
All placeholders in square brackets are variables. Mapped dynamically are [LEAD_NAME], [COMPANY_NAME], [BUDGET_INFO], [NEED_DESCRIPTION], [TIMING_INFO] and [ADDITIONAL_INFO], as described in the data flow. Not dynamic are [BUDGET_CRITERIA], [NEED_CRITERIA] and [TIMING_CRITERIA]. These are the company specific scoring criteria that you define manually before activation. The square brackets in the JSON format are format instructions to the language model and stay unchanged. List all placeholders that require manual configuration separately at the end.
SELF CHECK BEFORE COMPLETION
Check, before you report the workflow as finished:
1. Are all steps connected correctly, including the duplicate check before the CRM entry?
2. Is the scoring prompt inserted in full and unchanged, and are all form fields mapped to the placeholders?
3. Is the JSON validated, and does the routing work (A and B immediately, C and D into nurturing)?
4. Are all points listed that require manual configuration: the form webhook with required fields, CRM access with stage mapping, sales routing, [BUDGET_CRITERIA], [NEED_CRITERIA], [TIMING_CRITERIA] and the LLM access?
Working prompts
Lead scoring
Used in step 2 to assess budget, need and timing of the incoming lead and assign an A to D score with a rationale.
You are an assistant for lead qualification. Analyze the following lead data and assess the lead quality:
Lead information:
- Name: [LEAD_NAME]
- Company: [COMPANY_NAME]
- Budget: [BUDGET_INFO]
- Need/requirement: [NEED_DESCRIPTION]
- Timing/timeframe: [TIMING_INFO]
- Additional information: [ADDITIONAL_INFO]
Your task:
Assess the lead against the following criteria and assign an overall score.
1. Budget fit (0 to 10 points): [BUDGET_CRITERIA]
2. Need fit (0 to 10 points): [NEED_CRITERIA]
3. Timing fit (0 to 10 points): [TIMING_CRITERIA]
Scoring scheme:
- A lead (25 to 30 points): high priority, immediate contact
- B lead (18 to 24 points): good quality, contact within 24 hours
- C lead (10 to 17 points): medium quality, nurturing required
- D lead (0 to 9 points): low quality, automatic follow-up
Respond exclusively in the following JSON format:
{
"total_score": [total points],
"scoring_category": "[A/B/C/D]",
"budget_points": [0 to 10],
"need_points": [0 to 10],
"timing_points": [0 to 10],
"qualification_status": "[Qualified/Conditionally qualified/Not qualified]",
"recommended_action": "[concrete recommended action]",
"summary": "[2 to 3 sentences justifying the score]"
}
JSON export
Setup
Two routes lead to the same result. Pick the one that fits your setup.
Suitable when you want to build the workflow in your own automation tool.
Open your workflow tool, for example n8n, Make, Zapier or Langdock.
Create a new workflow and give it a name.
Choose the AI-assisted build option if your tool offers one.
Paste the builder prompt from above in full.
Check the result against the step sequence in this document, then add your access credentials and channels.
Suitable when you want to take the workflow over unchanged.
Download the JSON file from the JSON export section.
Open the workflow area of your tool and create a new workflow.
Give it a name and confirm.
Open the menu on the workflow name and choose to import a JSON file.
Upload the file, then add your own access credentials.
Run a test before you activate the workflow.
Last reviewed:
In the workshop this becomes your method.
A single prompt becomes a repeatable method. We show that in the workshop From Prompt to Method.
