AI Sales Academy
en

A Prospect Brief generated automatically

Automation 11 min
Business problem

Research before contact is 10-20 minutes per lead. At 20 contacts a day that's 4-7 hours. Most salespeople skip research and write 'Hi John' — losing the personalization that drives replies. Research automation brings the quality back without the time cost.

Proven method

The research workflow in 7 steps:

  1. NEW LEAD: a trigger (a lead after qualification),
  2. COMPANY RESEARCH: AI analyzes the website, description, business model (what they do, for whom),
  3. RECENT NEWS: AI searches news (last 90 days) — but always with a link to the source for verification,
  4. TECHNOLOGY: AI identifies the stack (what they use — from the website analysis and public data),
  5. HIRING: AI checks job posts (is sales/ops growing — a signal),
  6. BUYING SIGNALS: AI combines the above into signals (hiring + a new branch + a CEO change = a window),
  7. AI SUMMARY: a 3-5 sentence summary: the company's situation, 2-3 pain hypotheses, a recommended personalization angle + all facts with links. Output: a Prospect Brief (one card). Rules:
  8. AI generates a brief with facts LABELED with a source (so a human can verify),
  9. hallucinations: every fact in the brief has a link or a 'hypothesis' label (AI doesn't give news without a source),
  10. the brief is 'ready for verification', not 'ready to send' — a human checks the 1-2 most important facts.
Market example
BAD
Manual, inconsistent execution or chaotic automation lacking data validation.
GOOD
A Prospect Brief for 'MebelSoft': COMPANY: furniture e-commerce, 40 people, Shopify, sells to PL/EU. NEWS: 'expanded the warehouse in 2026' (article link), 'new office line' (link). TECH: Shopify, Klaviyo, no chat (from the website analysis). HIRING: 2 customer service posts (link) — inquiry volume is growing. SIGNALS: service growing without tools = a potential pain, no chat. SUMMARY: 'A growing e-commerce company, hiring for service, inquiry handling without automation — a good candidate for a contact about customer service systems. Angle: growth + no tools'. The salesperson verifies: the website + 1 job post. All in 3 minutes instead of 20.
Ready-to-use prompt 
You are an automatic research architect.

Context: My sources: [WEBSITE, NEWS, LINKEDIN, JOB POSTS — what I can search automatically]. My product: [PRODUCT]. Tools: [AI + MAKE/N8N + SOURCES].

Task: Design the research workflow:
1. A schema: New Lead → Company Research → News → Technology → Hiring → Buying Signals → AI Summary
2. For each step: what data, from where (API/source), how AI processes it
3. A Prospect Brief: the card structure (sections + fields) — what goes into the brief
4. The sources rule: how AI labels facts (link/hypothesis) and how I avoid hallucinations
5. Buying signals: how AI recognizes signals from data (rules)
6. Human-in-the-loop: what a human verifies in the brief (min 1-2 facts)
7. Integration: how the brief goes to the CRM/campaign (structured output)

Format: 7 sections.
Model rekomendowany: Claude 3.5 Sonnet / GPT-4o Copy and paste into ChatGPT, Claude or Perplexity
ALTERNATYWNE SCENARIUSZE

Variations & alternative scenarios

Light version: research in 3 steps (company, news, signals) without hiring/tech.

Full version: research from many sources + fact verification.

Practical exercise

Design the Prospect Brief structure (sections + fields). List 3 sources you can search automatically. Write the rule: how you label facts (link vs hypothesis) to avoid hallucinations.

Zapisano w przeglądarce

Steps to take

0/4