Your AI agent just
learned
how to prospect.

Goal: a verified, deduped lead list, one decision-maker per company. Target: exactly 25 people whose current title is one of ["VP Marketing","Head of Marketing","CMO"] (fall back to "Marketing Director" only if none exist at that company), at B2B SaaS companies, HQ country = United States, headcount 50 to 200, founded 2015 or later. Process: 1. Build a 30-company shortlist (over-fetch to cover misses) and print it as a numbered list with company + employee count. STOP and wait for my "go". 2. After "go", pick the single best-fit person per company (one per company, no duplicates). 3. Enrich with Scalelist: verified work email + direct mobile. Cap spend at 60 credits; if you'd exceed it, stop and ask. Output table, columns in this exact order: Rank | First name | Last name | Title | Company | Employees | HQ city | Website | LinkedIn URL | Email (lowercase) | Email status (valid/risky) | Mobile (E.164). Rules: include ONLY rows with a valid email; drop risky/not-found; sort by Employees descending; never fabricate a value (leave blank + note). Finish with: rows returned, emails found, mobiles found, credits used.

I'm attaching a CSV. Identifier per row is either (Full name + Company domain) OR a LinkedIn URL. Process: 1. Tell me the row count detected and which identifier you'll use. If >200 rows, process in batches of 100 and confirm after each batch. 2. Enrich every row with Scalelist: verified work email + direct mobile. Verify each email and label it exactly valid / risky / not_found. Output: return my ORIGINAL columns and row order unchanged, then append, in this order: Email (lowercase) | Email_status | Mobile (E.164) | Mobile_status (found/not_found) | Reachable (Yes if Email_status=valid OR Mobile_status=found, else No) | Note (short reason when not reachable). Then print a summary: rows processed, valid emails, risky, not_found, mobiles found, reachable %, credits used. Misses cost 0 credits, never delete unreachable rows, only flag them. Finally export as "enriched_<original-filename>.csv".

Write a personalized cold email for each of these 10 enriched contacts. My context: I sell [product] to [audience]. Primary outcome: [metric/result]. Proof point: [1 customer + result]. CTA: interest-based, not a meeting ask. For each contact, mine their Scalelist profile for ONE concrete, verifiable hook, in this priority: (1) job change in last 6 months, (2) company signal [hiring/funding/launch], (3) specific role scope. If none exist, mark "needs manual review" and skip, never fabricate a hook. Hard rules: subject ≤ 42 characters, no emoji; body 50 to 90 words; ≤ 9th-grade reading level; plain text; exactly one CTA phrased as a question; banned phrases: "hope this finds you well", "quick question", "I noticed you're the [title]", "synergy", "circle back", "game-changer". Output table: # | Name | Company | Hook type used | Subject | Body | Word count.

Map the GTM org at [company.com] so I know exactly who to open with. My offer: [what I sell + outcome + who owns it]. Process: 1. Use Scalelist to find current employees in Sales, Marketing, and RevOps/Operations at this domain. 2. Group by function; within each, rank by seniority: C-level > VP > Senior Director > Director > Senior Manager > Manager > IC. 3. Recommend a Primary and Backup first-contact for my offer, each with a one-line reason tied to budget or problem ownership (not just seniority). 4. Enrich ONLY the Primary and Backup with verified email + mobile (max 2 credits). Output: (A) org chart as an indented list grouped by function, each person = Name, Title, tier; (B) "Contact first" block with Primary {name, title, why, email, mobile} and Backup {same}. Edge case: if the company has <10 employees or no GTM titles, say so and name the founder/CEO as the entry point. End with credits used.

Build a 5-touch outbound sequence for these 20 enriched leads. I sell [product]; outcome [result]; proof [stat]; CTA style = soft/interest-based. Exact cadence (business days): Day 1 Email, Day 2 LinkedIn DM, Day 4 Email, Day 6 LinkedIn DM, Day 9 Email (breakup). Angle per touch (no repeats): T1 problem hook → T2 one-line credibility → T3 new proof/use case → T4 light social-proof nudge → T5 breakup with an easy out. Personalize every touch from each lead's Scalelist profile (role, company, signal); the T1 opening line must be unique per lead. Emails 50 to 90 words, DMs ≤ 45 words, one CTA each. Output: one master table. Lead | Company | Touch | Channel | Send day | Subject (emails only) | Message | Word count. Then: (a) the 1 variable I should A/B test first and why; (b) any lead with too little data flagged "generic, needs review" rather than guessed.

I have a call with [name] at [company] in 10 minutes. Pull their profile + company via Scalelist and brief me in EXACTLY this structure, total ≤ 220 words, no fluff: 1) Who they are, title, seniority tier, time in current role, previous role/company (1 line). 2) Company, what they sell, segment, employee band, and 1 recent signal with rough date. 3) Their likely top 2 priorities + 1 pain for this exact seat. 4) Fit, one sentence mapping [my product] to point 3. 5) Opener, one tailored line referencing point 1 or 2 (not generic). 6) Discovery, 3 questions ordered easy → hard, each ≤ 20 words. 7) Landmine, 1 objection/sensitivity common to this persona + a one-line counter. If any field is unknown, write "unknown, not in data" rather than guessing. Don't enrich email/mobile.

Push these 20 enriched contacts to my [HubSpot/Salesforce/Pipedrive]. Process (do NOT write anything until step 3 is approved): 1. Dedupe: match existing records on email (exact, case-insensitive); if no email, match on (Last name + Company domain). 2. Show me a preview: New (n), Duplicate (n, with matched record name+ID), Conflict (n, where my data differs). Wait for my approval. 3. On approval: create New; for Duplicates only fill BLANK fields (never overwrite); skip Conflicts and list them. Field mapping: First name, Last name, Title, Company, Email (lowercase), Mobile (E.164), LinkedIn URL, Lead source = "Scalelist MCP". Apply tag/list "Scalelist. Q2 2026". Add a note per record: "[industry], [employees], fit: [one line]". Output: summary table. Created | Updated (blanks filled) | Skipped duplicate | Skipped conflict | Failed (with reason).

Source 15 candidates for: [Senior Backend Engineer]. Must-have: [Python] + [5+ yrs backend]. Nice-to-have: [AWS, Kubernetes]. Location: [remote, France or CET ±2h]. Hard exclude: anyone currently at [our company] or [competitors]. Process: 1. Build a 20-name shortlist (over-fetch); print Name | Current title | Company | Location; wait for my approval. 2. On approval, enrich each with Scalelist (verified email + mobile; cap 30 credits). 3. Score fit 1 to 3 (1 = best) against must-haves only; nice-to-haves break ties. Output table, sorted Tier 1→3 then years desc: Tier | Name | Current title | Company | Location | Years relevant | Must-haves met (y/n) | Why (≤15 words) | Email | Email status | Mobile (E.164). Flag candidates with an "open to work"/recent-departure signal as "⚡ priority". End with count per tier + credits used. Never include an excluded company.

Seed (a customer who converted last month): [LinkedIn URL or Name + Company]. Process: 1. Profile the seed + company via Scalelist and output a 1-line ICP: industry | employee band | role+seniority | region | 1 to 2 buying signals. 2. Wait for my "go" / edits on that ICP. 3. Find 10 lookalike contacts at OTHER companies (exclude the seed's company and [my company]); one per company. 4. Score each 0 to 100 on ICP fit, weighting: role/seniority 40, industry 25, company size 20, region 15. Enrich ONLY contacts scoring ≥ 60 (top 5 max) with verified email + mobile. Output: (A) the final ICP line; (B) ranked table sorted by score desc: Rank | Name | Title | Company | Employees | Region | Match score | Score reason (≤15 words) | Email | Mobile (E.164). Rules: don't enrich anyone < 60; if fewer than 10 hit the bar, return only those and say why. End with credits used.