"AI-personalized" has become one of the most overloaded phrases in cold outreach. In practice, it covers two very different things: tools that use AI to fill in merge fields more fluently, and tools that use AI to actually research the recipient before writing anything. The first is a nicer template. The second changes what a cold email even is.

Merge fields are not personalization

{{firstName}}, {{company}}, {{title}} — swapping these into a fixed template is table stakes, not differentiation. Every recipient of a merge-field campaign is reading the exact same email with a few nouns changed, and both people and spam classifiers have gotten very good at recognizing the pattern. It scales perfectly and persuades nobody.

Real personalization requires research, not just data

The version that actually moves reply rates is structurally different: the AI researches the specific company before writing a single word — recent news, what the company's website says about itself, signals about what they're likely working on right now — and writes one short, specific paragraph grounded in that research. The rest of the email (structure, offer, call to action) stays consistent across the sequence; only that one paragraph is genuinely written fresh for the person receiving it.

The tell: does the claim survive scrutiny?

A useful test for any "personalized" line in a cold email: could it have been written without knowing anything about this specific company? "I noticed you're focused on growth" survives being generic — it's true of almost every company that gets a cold email. "Saw you just opened a Bangalore office" doesn't survive being generic; it's either true and researched, or it's false and will be noticed as false within one sentence.

• Generic personalization: fills in a name and company from your lead data.
• Real personalization: writes a claim that's specific enough to be wrong if the AI didn't actually look anything up.
• The second kind is measurably harder to fake — which is exactly why it works better.

Why this matters more as AI writing gets more common

As AI-written cold email becomes the norm rather than the exception, the bar for what reads as "clearly automated" keeps rising, and recipients get faster at recognizing template-with-a-name-swapped patterns. The gap between merge-field automation and research-grounded writing isn't going to close — it's going to widen, because the first kind gets easier to spot the more common AI writing becomes, and the second kind gets easier to produce well as the underlying research and writing models keep improving. Betting on genuine per-prospect research now is the version of "AI personalization" that keeps working as everyone else's stops.