AI Sales Academy

Personalization vs customization and the 5 levels

Prospecting 12 min
Business problem

Most 'personalization' is customization: inserting a name into a template. It gets no replies because it says nothing about the customer. Real personalization requires research and work — and you can't do it by hand for 100 leads. You need a process that turns research into level 3-5 personalization.

Proven method

The 5 personalization levels: Level 0 — 'Hi John' (customization: a name, nothing more), Level 1 — company ('I see you work at X'), Level 2 — position ('as Head of Sales you manage the pipeline'), Level 3 — problem ('I see you lose follow-ups between the phone and Excel'), Level 4 — trigger ('you're hiring 3 salespeople'), Level 5 — trigger + business context ('you're hiring 3 salespeople and calls are handled over one phone — looks like the new people get more work, not a system'). Goal: at least Level 3 for everyone, Level 4-5 for Priority 1. AI workflow: research (levels 1-2) → AI identifies the problem (3) and trigger (4) → combines them into business context (5) → generates 3 opener variants. The rule: personalization MUST be real — a research fact that matters for the customer's business.

Market example
BAD
Unstructured prospecting without ICP filters and generic outreach to random accounts.
GOOD
The difference is visible in the opener: L0 'Hi John,' L1 'I see you work at company X' L2 'As Head of Sales you own the pipeline' L3 'Company X loses leads between the phone and a spreadsheet — that's common when sales grows' L4 'You're hiring 3 salespeople — your handling scale is growing' L5 'You're hiring 3 salespeople and opened a branch in Gdansk, but leads still land on one phone — the new people have nowhere to log conversations'. L5 is an opener people reply to.
Ready-to-use prompt
You are a B2B personalization specialist.

Context: I sell [PRODUCT]. Company research: [RESEARCH FACTS]. Person research: [ROLE, SIGNALS].

Task: Generate 5 personalization levels for the first email sentence:
Level 0: name
Level 1: company
Level 2: position
Level 3: problem (from research)
Level 4: trigger (from research)
Level 5: trigger + business context (combine 2 facts, show the consequence for the company)
For L3-L5 state WHICH research fact you use (so it's real personalization).

Format: 5 rows.
Model rekomendowany: Claude 3.5 Sonnet / GPT-4o Copy and paste into ChatGPT, Claude or Perplexity
ALTERNATYWNE SCENARIUSZE

Variations & alternative scenarios

Email vs LinkedIn version: show which level works better in each channel.

Scale version: 3 L3-L5 variants for 5 leads at once (one per lead).

Practical exercise

Take 3 Priority 1 leads. For each write openers at levels 0-5. For L3-L5 write which fact you use. Pick the best for each lead.

Zapisano w przeglądarce

Steps to take

0/4