When sales directors are asked where generative AI will have the most impact, prospecting comes out on top.
In this new article, you'll learn how to:
define a highly targeted ICP with AI,
conduct contextual monitoring with ChatGPT,
generate enriched lead lists with Freckle,
write email sequences that convert.
All this, without coding, using accessible tools (free or freemium).
1/ Define a precise ICP with the "context file"
A clear ICP (Ideal Customer Profile) = less cold outreach, higher response rates.
Quick method:
Create a context file (PDF or doc) containing: pitch, offers, figures, personas, client cases.
Import it into ChatGPT (or Claude/Gemini).
Ask: "Act as a Go-to-Market expert. Analyze this file and provide me with 2–3 example ICPs with key function, sector, purchase signals."
Side School Tip:
With ChatGPT Plus, activate memory: add "Put in memory" to save your pitch once and for all.
2/ Go further with in-depth research
The "deep research" option allows you not only to define your ICPs more precisely but also to conduct contextual research on your key accounts.
Model prompt:
"What AI/GenAI initiatives has Carrefour launched in 2024–2025? Summarize and cite your sources."
In a few minutes, you get: pilot projects, technology partners, announced budget. Ideal for breaking the ice and customizing your pitch.
3/ Generate & enrich your leads in natural language
To build a truly qualified prospect list, combine three approaches:
Free search with Exa
Simply type your query in natural language: "French SaaS start-ups that have raised more than one million euros in 2024." In seconds, Exa returns a ready-to-use file with company names, URLs, and initial contacts detected.
Hybrid filtering in Clay
Start with "classic" filtering: sector, size, country. Once the list is obtained, refine it using AI queries: Clay recalculates the list in real time.
Hyper-enrichment in Freckle
Import a simple CSV containing first name, last name, company. Then add smart columns — professional email address, LinkedIn profile, TikTok account, etc. For each column, Freckle searches for the information on the web and fills in the sheet.
4/ Write ultra-personalized sequences
V7 Go is a spreadsheet that integrates AI, allowing the stacking of multiple models.
Step-by-step workflow:
1. Column A: Exported LinkedIn CV (PDF).
2. Column B: `First Name` → extraction GPT‐4o.
3. Column C: `AI Pain Points` → CV analysis.
4. Column D: `Job` → single-select "Sales / HR".
5. Columns E-F: Email #1 and Email #2 drafted by Claude, tailored to the role and detected needs.
Result: custom messages, sent with one click (or via Zapier/API to your CRM).
5/ Frequently Asked Questions
Do you need all the tools?
No: choose the one that integrates with your CRM and covers 80% of your cases.
Are my prospect data kept private?
Use "Enterprise / Pro" versions that exclude your data from model training. See "AI & Data" Side School report.
Can you go as far as an autonomous AI agent?
Yes, via API (VC → Sendgrid, Lemlist, etc.) + control loops. But always measure the ROI before scaling.
Find more answers in our AI reports.
Auteur :
Biographie
Directeur Associé chez Side School. Ben Issen était précédemment fondateur de Supercreative où il a créé plusieurs outils IA à destination des freelances.