Lead Qualification Prompts Your CRM Can Actually Use

Most CRMs have lead scoring. Most lead scoring is useless. Static fields, point values nobody updates, scoring rules from 2021 still firing in 2026.

These prompts replace the static system with something that reads what the lead actually said and scores accordingly.

Prompt 1: Intent scoring from form text

``` Score lead intent based on what they wrote in an inbound inquiry.

Lead's inquiry text: {TEXT} Their stated company: {COMPANY} Their stated role: {ROLE}

Score 1-10 on: - Specificity (are they describing a real problem or asking general questions?) - Urgency (do they have a deadline or pressure?) - Authority (does their role suggest decision power?) - Fit (does their problem match what we solve?)

For each score, quote the specific phrase from their inquiry that drove the score.

End with: total score / 40, recommended priority (hot / warm / qualify further / not a fit). ```

The quote requirement keeps the AI honest. Salespeople trust scores they can audit.

Prompt 2: Disqualifier scanner

``` Scan a lead's inquiry for disqualifiers.

Inquiry: {TEXT} Our disqualification criteria: {LIST}

Common disqualifiers: - Company size below our minimum - Industry we don't serve - Asking for something we don't offer - Vendor research signals (asking for case studies before discovery) - Compliance regime we don't support - Geographic requirement we can't meet

For each disqualifier found, quote the trigger phrase and explain.

If no disqualifiers, return "no disqualifiers detected." ```

The disqualifier scan is the fastest filter. Catches the 30% of leads that shouldn't enter the pipeline.

Prompt 3: Industry-specific qualification

``` Qualify this lead against criteria specific to {INDUSTRY}.

Lead's inquiry: {TEXT} Lead's company: {COMPANY} Industry-specific qualifications: {CRITERIA}

For a {INDUSTRY} lead, the things that matter most are: {DETAILED_CRITERIA_LIST}

Score each criterion (Y/N/Unclear). For Unclear, suggest the specific question to ask in discovery to clarify.

End with: criteria met / total, recommended next step. ```

Industry-specific qualification is where generic CRM scoring fails. A wealth manager lead and a CPA lead need different criteria.

Prompt 4: The "where are they in their journey" classifier

``` Classify the buyer journey stage of an inbound lead.

Lead's inquiry: {TEXT} What they've engaged with on our site: {PAGES_VISITED}

Classify as one of: - Awareness (just learning the space) - Consideration (comparing options) - Decision (ready to engage, picking a vendor) - Implementation (already bought somewhere, looking for help) - Not yet (no real need, just curious)

Quote the phrase that drove the classification.

Recommended outreach approach for this stage: [educational content / discovery call / proposal / implementation scoping / nurture]. ```

Different stages get different first responses. AI-classified stage means the response can be auto-routed.

Prompt 5: The unstructured-data scanner

For when the lead comes through with a long email instead of a form.

``` Extract structured fields from an unstructured lead email.

Email: {EMAIL}

Extract: - Their company (if mentioned) - Their role (if mentioned) - Company size hint (if any) - The specific problem they're describing - Their timeline (if any) - Their budget signal (if any) - Anyone else they mentioned by name - Any current vendor or tool they mentioned

If a field isn't in the email, return "not specified". Do NOT infer or guess.

Return as JSON for direct CRM ingestion. ```

The "do not infer or guess" instruction is essential. Without it, the AI fills gaps with plausible-but-wrong data that pollutes the CRM forever.

My actual setup

Inbound form → webhook to n8n → run prompts 1, 2, 4, 5 in parallel → results merged → CRM entry created with: - Auto-scored intent (1-40) - Disqualifier flags (if any) - Journey stage - Extracted structured data

The salesperson sees the score, the quotes that drove it, and the structured summary. They don't read the raw email first. They read the analysis, then the raw email if they want context.

Time per lead before: 5-10 minutes of triage. Time per lead after: 30 seconds.

What I'd change about traditional lead scoring

Traditional lead scoring is: - Points for company size - Points for industry - Points for role title - Points for page views

These signals are weak. They tell you nothing about intent.

The right signals are: - The specificity of what they wrote - The urgency phrases they used - The disqualifiers they revealed - The fit between their problem and what you sell

These are not in the CRM as structured fields. They're in the text. Prompts can extract them. Traditional scoring can't.

What I'd build first

Prompt 2 (disqualifier scanner). It's the highest-leverage one because it stops bad leads from entering the pipeline.

Run it on every inbound. Auto-tag the CRM record. Salespeople prioritize untagged leads. The disqualified ones get a polite auto-decline with a referral.

The disqualified leads still get treated well. The salespeople stop wasting time. Win-win.

What changes the math

When you replace static scoring with prompt-based scoring, the salespeople start trusting the score. They use it. They stop ignoring the CRM.

That trust is the actual win. The technology is incidental.