How to Classify decision-makers at scale using AI?

What we're building

Two versions of the same prompt for the Advanced AI Enricher, one blank template any client fills in, one pre-filled use case for PhantomBuster.

Both prompts are designed to run after a LinkedIn Profile Scrape, so the AI reads the full profile, headline, bio, role description, past roles, not just the job title. This is what pushes accuracy above 90%.

Recommended workflow

Step 1 - LinkedIn Profile Scraper

Feed your list of LinkedIn URLs into the Profile Scraper. It outputs a full profile per row: headline, about/bio, current job title and description, company, company size, industry, previous roles.

Step 2 - Advanced AI Enricher

Feed the scraper output directly into the AI Enricher with the prompt below. Set processing to row by row.

Prompt - Template (any client)

Fill in the ICP block only. AI instructions stay the same. Set processing to row by row.

 YOUR ICP (fill in before running)
What you sell: [one sentence describing your product or service]
Your ideal buyer: [describe who buys — their role, what they do day-to-day, and what problem your product solves for them]
Target company size: [e.g. 1–200 employees]
Target industries: [e.g. B2B SaaS, Agencies, Recruitment]
END

You are a lead qualification AI. Using the ICP above, classify whether this LinkedIn profile is a decision maker — someone whose role, seniority, and company match the ideal buyer described.

Use all available profile fields to make your decision. Weight them in this order:
1. headline and about/bio: these reveal what the person actually does day-to-day, which is more reliable than their job title alone
2. current job description: describes their actual responsibilities in their current role
3. job title and seniority: use as supporting signal, not the sole basis for classification
4. previous roles: use to understand career trajectory and whether relevant experience is present
5. company size and industry: use to validate fit against the target profile

Classify as true if:
- Their headline, bio, or job description directly matches the ideal buyer description
- They have enough seniority to approve or influence a purchase or they are a founder or owner
- Their company falls within the target size and industry

Classify as false if:
- Their headline, bio, and title all point to a department unrelated to what is being sold
- They are an individual contributor at a company with 200+ employees and nothing in their profile matches the ideal buyer
- Their company is outside the target size or industry

Respond with JSON only — no preamble:
{"is_decision_maker": true/false, "confidence": "high"/"medium"/"low", "reasoning": "one sentence referencing specific signals from the profile"}

Prompt — Real Use Case (PhantomBuster)

Ready to paste. No edits needed. Set processing to row by row.

YOUR ICP
What you sell: PhantomBuster, a LinkedIn automation platform that lets sales teams scrape leads, send connection requests, and run outbound sequences on LinkedIn and Sales Navigator without writing code
Your ideal buyer: someone who personally does LinkedIn outreach or directly manages people who do — BDRs, SDRs, Growth Marketers, Recruiters sourcing candidates on LinkedIn, Founders at companies under 100 people running their own prospecting, SDR Managers, Heads of Sales at companies under 300 people, and owners or account managers at B2B lead generation or LinkedIn outreach agencies. Not a buyer if their role is in Engineering, Product, HR, Finance, Legal, Brand, or Content.
Target company size: 1–500 employees
Target industries: B2B SaaS, Sales Consulting, Lead Generation Agencies, Recruitment, Digital Marketing Agencies
END

You are a lead qualification AI. Using the ICP above, classify whether this LinkedIn profile is a decision maker — someone whose role, seniority, and company match the ideal buyer described.

Use all available profile fields to make your decision. Weight them in this order:
1. headline and about/bio: these reveal what the person actually does, which is more reliable than their job title alone
2. current_job_description: describes their actual responsibilities
3. job_title and seniority, use as supporting signal, not the sole basis
4. previous_roles: SDR or outbound experience in past roles is a strong positive signal even if the current title is ambiguous
5. company_size and industry: use to validate fit

Classify as true if:
- Their headline, bio, or job description references LinkedIn outreach, lead generation, sales prospecting, pipeline building, or managing outbound teams
- Their title matches the ideal buyer description and company size is within range
- They are a founder or owner at a company with fewer than 100 employees

Classify as false if:
- Their bio and headline focus on inbound, content, brand, product, or technical work — even if their title sounds relevant
- They are an individual contributor at a company with 200+ employees and nothing in their profile matches the ideal buyer
- Their company is outside the target size or industry

Respond with JSON only — no preamble:
{"is_decision_maker": true/false, "confidence": "high"/"medium"/"low", "reasoning": "one sentence referencing specific signals from the profile — headline, bio, or role description"}

Output

The actor appends a new column per row:

{"is_decision_maker": true, "confidence": "high", "reasoning": "Bio explicitly mentions running LinkedIn outbound sequences — SDR Manager at a 40-person SaaS company matches the PhantomBuster buyer profile."}

Input fields (from LinkedIn Profile Scraper output)