6 Ways I'm Using AI to Write Cold Emails That Actually Get Replies
Troy Aitken shares 6 concrete AI tactics his team uses right now to write cold emails that feel researched, relevant, and personal.
Most people using AI for cold email are doing it wrong. They're asking it to write a generic opener, slapping a name on it, and calling it personalization. What I'm doing is different: I'm using AI to do the research work that makes an email feel like it was written by someone who actually gave a damn.
Here are the six ways my team is running AI in our cold email process right now.
1. Call Out Specific Competitors in Their Vertical
This is one of the best-performing tactics we're running for a content writing agency client right now. I task GPT to research the prospect's company, then identify a handful of direct competitors operating in the same vertical. When you name a competitor in an email, it signals immediately that you understand their world. You're not pitching into a void. You know who they're up against.
We're also using this same competitive research on the SEO and backlink side of things for that client, so the AI output does double duty.
2. Map the Right Case Study to the Right Prospect
This one runs inside Clay. The goal is straightforward: figure out which of our existing customer case studies is the closest match to the prospect we're reaching out to. Not just industry, but situation, pain point, and outcome.
When you drop a case study into a cold email, it needs to feel like you picked it for them, not pulled it from a random rotation. AI helps us make that match at scale. The result is a proof point that lands because it actually mirrors what the prospect is dealing with.
3. Name the Industry They're In
This sounds obvious, but most cold emails skip it entirely or get it wrong. I make sure the email explicitly calls out the vertical the prospect operates in. It's not a flashy move. It just demonstrates that you know exactly who you're talking to and that you haven't sent the same email to a plumber and a SaaS founder.
4. Reference What's on Their Website
This is where the research gets specific. I'm using AI to pull details directly from the prospect's website: recent promotions they're running, services they list, products in their catalog. Then those details show up in the email.
The effect is simple but powerful. The prospect reads it and thinks, "Someone actually looked at my site." That perception of effort, whether it came from a human or a well-prompted AI, changes how they receive the message.
📥 Cold Email Swipe File
Steal the cold email templates our clients used to generate $8M+ in revenue.
5. Scrape and Call Out Specific Products (Especially for Ecom)
If you're running cold email campaigns targeting e-commerce brands, this is the one to prioritize. I'll script the prospect's website to pull a specific product, then reference it directly in the email copy.
It works because it's concrete. You're not saying "I noticed you sell products online." You're saying "I saw you're running [specific product]." That level of detail is what separates an email that gets deleted in two seconds from one that gets a reply.
6. Combine All of It Into One Cohesive Email
Each of the five tactics above is useful on its own. But the real output is an email that weaves all of them together: a competitor callout, a matched case study, an industry reference, a website detail, and a product mention, all in one message that reads like a human did serious homework.
AI makes this possible at volume. The key is building the prompts and the Clay workflows so the research feeds directly into the copy. When it's set up right, every email in a sequence feels like it was crafted for that one person.
Key Takeaways
Use AI to research competitors in a prospect's vertical and name them in the email
Match your case studies to prospects based on situation and industry, not just gut feel
Always name the specific vertical you're targeting, it's a basic signal that you know your audience
Pull real details from the prospect's website (promotions, services, products) and put them in the copy
For e-commerce prospects, scraping and referencing a specific product is one of the highest-signal moves you can make
The goal isn't one tactic in isolation, it's combining all of them into a single email that feels genuinely researched
Frequently Asked Questions
What tools are you using to run these AI research tactics? The main ones I'm using are GPT for competitor research and website analysis, and Clay for mapping case studies to the right prospects at scale.
Does calling out a competitor in a cold email actually work? Yes, and it's one of our best-performing tactics right now for a content writing agency client. Naming a competitor signals that you understand the prospect's market, which immediately makes the email more credible.
Why is referencing a specific product so effective for e-commerce outreach? Because it's concrete proof that someone looked at their business. Even if the research was AI-assisted, the prospect sees a specific product name and concludes that real effort went into the email. That perception drives replies.
Do I need Clay to use these tactics? Clay is what we use for case study mapping, but the core idea, matching your proof points to your prospect's situation, can be done with other enrichment tools or even manually for smaller lists. Clay just makes it scalable.
This is one piece. The full system is the 2-hour masterclass.
Drop your email and get the full 15-chapter system — offer, list, copy, sequences, deliverability — start to finish.
More on copywriting.
Your pipeline, rebuilt.
20-minute strategy call. We'll audit your ICP, show you which signals we'd track, and map out exactly what the first 120 days would look like. No commitment, no pressure, no pitch deck.
