Key takeaways
✔️ Bad personalisation is worse than none buyers spot fake facts instantly
✔️ Good AI is just a draft - real context makes it relevant
✔️ Use signals: site visits, real convos, multi-threaded outreach
✔️ Personalisation should feel natural, not forced
In this GTM Insider Podcast we caught up with, Rob Harlow, Co-founder and Chief Innovation Officer at Sopro. Rob breaks down what real personalisation looks like and what fails instantly.
We all know that feeling when an email lands in your inbox and you clock in two seconds flat: This is AI. The problem? Bad personalisation tries to impress with irrelevant details — your university, your last LinkedIn post but adds no real value.
“Within the first kind of two seconds you’ve made the decision: this is AI. And I think literally every single day people get more and more sensitive to that.”
— Rob Harlow, Sopro
At Sopro, Rob’s team has sent more than 80 million emails, so they know what works. They do use AI, but with guardrails. An AI draft is just the start. What really moves the needle is layering in live context: your conversations with that company, who you’ve spoken to already, signals like website visits or multi-threaded contacts.
It’s not about name-dropping random facts - it’s about showing you actually understand their business.
FAQs
➜ What is bad personalisation?
Any random fact that adds zero value, like dropping someone’s university name for no reason.
➜ Is AI enough for personalisation?
No - AI should help draft, but humans need to add real, live context.
➜ How does multi-threading help?
If you’re talking to the CEO and the CTO, referencing both makes your message feel human and relevant.