Most cold email “personalisation” is the same template with a first name swapped in. Recipients recognise it instantly and it has no measurable impact on reply rate. Real personalisation requires a specific, true, relevant detail about the recipient that could not have been sent to anyone else. This guide covers what actually moves the needle, what is performative, and how to use AI to scale genuine personalisation.
Something specific and timely happened - they raised a round, posted a job, published an article, gave a talk, or made a statement in public. Referencing this says “I am paying attention to you specifically right now.” Highest-converting because it creates genuine relevance and timeliness.
A specific detail about their company - tech stack, recent product launch, company size milestone, geographic expansion, competitor they are mentioned alongside, or content they published. Shows you researched the company, not just filled a template.
Content specific to their role and seniority - not their name, but their situation as a [VP of Sales at a 50-person B2B SaaS company]. Can be templated but must be tightly segmented. Generic ICP copy sent to a mis-matched segment performs worse than no personalisation.
{{first_name}} and {{company}} are baseline expectations, not personalisation. Every recipient knows the name is merged from a field. It does not reduce friction or create relevance on its own. Only valuable as a carrier for higher-level content.
“I love what you’re doing” or “Really impressive work at {{company}}.” Recognised instantly as performative. Has no measurable positive impact and may reduce trust because recipients know it is fake.
Feed a list into Clay. Use Claygent (AI browsing agent) to research each contact - extract their last LinkedIn post topic, most recent company news, or a specific claim from their website. Output a custom first-line variable. Scale: 500-2,000 contacts/day. Cost: ~$0.10-0.50 per contact in Clay credits + Claygent usage.
Build separate campaigns per trigger segment: Hiring SDRs, Raised Funding (last 90 days), New Job (last 60 days), Posted on LinkedIn (last 7 days). Each segment gets a template written for that trigger. Zero manual research per contact. Scale: unlimited. Cost: ~$0.01-0.05 per contact for trigger data.
For your top 20-50 target accounts, do manual research: read their content, understand their product, find specific details that justify reaching out. Write individually crafted emails. No AI. Scale: 5-15 emails/week per SDR. Appropriate when ACV exceeds $50K.
Writing a custom first sentence but using a generic value prop that could apply to any company. The opener creates credibility; the offer must convert it. Both must be relevant.
No amount of personalisation salvages an email to someone who has no reason to buy. Personalisation amplifies relevance - it cannot create it where none exists. Fix the list, not the copy.
Claygent and similar tools sometimes hallucinate details - citing funding that did not happen, misquoting a LinkedIn post, or attributing content to the wrong person. Verify AI-generated first lines on a sample before deploying at scale.
Referencing very specific personal details (a specific tweet from 2 years ago, a photo from a conference, a comment in a niche forum) crosses from personalised to uncomfortable. Professional context only.
If the first email has a strong personalised opener and steps 2-5 are generic, the reply rate on follow-ups drops steeply. Maintain relevance signals across the sequence, not just in email 1.
A perfectly personalised email to an invalid address is zero. Verify your list before investing in personalisation research. 100 free credits, no card required.
Strategy, writing, sequences, and lead generation
Continue through related topics