Saw their claims about AI-generated cold emails. Decided to test it myself with a real sequence for my audit tool.
I fed it:
* My ICP profile
* Three competitor landing pages
* Our core value prop (no fluff, just compliance checks)
What I got back was...generic. The "personalization" was just inserting company names and weak pain points scraped from the sites. The "performance predictions" feel like complete guesswork.
Key issues:
* No verifiable data on where the "predicted performance scores" come from. Their own training data? Not disclosed.
* The generated copy defaults to marketing hype. Had to force it to be more factual.
* The data handling terms are vague. What exactly are they doing with my ICP data, competitor info, and my own value propositions?
Has anyone else done a real, side-by-side test of the output against a human-written sequence? I'm looking for:
* Actual A/B test results from a live campaign
* Any transparency on the scoring algorithm
* How they handle the data you input
Yeah, the performance scoring part is what made me skeptical too. It feels like a black box. I tried a similar tool last month for outreach and got a "87% predicted open rate" with zero explanation.
Did you find any way to actually audit those predictions? I'm wondering if there's an API endpoint or something where you could ping their model with sample input to see the raw scoring factors. Might be hidden behind their main interface.
If they're using their own training data, that's a huge transparency issue. My guess is it's trained on generic marketing copy, not actual cold email reply data.