Hi there! I'm a security recruiter using ChatGPT to help write cold outreach messages. In the Cold Outreach Message column, I'm trying to prompt ChatGPT to generate personalized outreach that ties in each cybersecurity consultancy’s website/description with how my firm can support them. Right now, some outputs are solid, but others include fluff like “Hi there” or “Here’s a tailored message…” before getting to the body. In some cases, I’m also getting high-level overviews of the company rather than messaging that reflects our firm’s specific experience and how we can add value to the company’s mission. I’ll share a few examples and my clay table below to show the tone and structure I’m aiming for — would love help refining the prompt so I get more consistent, useful output.
To refine your ChatGPT prompt for cold outreach: 1. Establish clear context: "You are an expert at writing concise, personalized outreach messages that highlight specific value propositions." 2. Avoid mentioning "cold email" terminology in your prompt. 3. Create specific instructions: "Generate a personalized message that connects the prospect's cybersecurity focus with our firm's relevant experience. Skip generic intros like 'Hi there' and focus on our specific value proposition." 4. Use our write to table feature to integrate these personalized messages directly into your outreach workflow.
To refine your ChatGPT prompt for personalized outreach in your Clay table: 1. Start with a clear role instruction: "You are writing personalized cold outreach that connects the prospect's needs with our specific capabilities." 2. Set format requirements: "Skip generic intros like 'Hi there' and avoid company overviews. Start directly with a value proposition." 3. Include specific instructions about your firm's experience: "Reference our [specific expertise/case studies] and how they directly address the prospect's challenges." 4. Use our write to table feature to integrate these messages directly into your outreach workflow. 5. Test your prompt with a few examples and refine based on results.
examples: Saw your post about the open Senior AppSec Engineer role at FusionAuth — sounds like a critical hire, and I know how tough it can be to find strong security engineers with app-layer depth and developer empathy. I run a recruiting firm that’s helped teams hire senior ICs in Application Security — especially those who’ve built secure SDLC practices, run threat modeling, and worked closely with devs to embed security into CI/CD. If you’re open to it, I’d be happy to share a few candidates who are either Denver-based or excited to relocate for the right mission. Let me know if that’d be useful — happy to help however I can. Saw your post about the Senior Security Engineer role at REI — great mix of cloud posture (CSPM/CNAPP), WAF, and bot mitigation focus. It’s rare to see that blend of strategy and systems-level depth in a mission-driven org like yours. I run a recruiting firm that’s helped security teams hire senior ICs with hands-on experience in cloud hardening, bot defense, and large-scale web application protection — especially in retail and consumer tech environments. If helpful, I’d be happy to share a few US-based candidates who’ve done this at scale. Let me know if you'd be open to that — always happy to help. aw your post about the Security Engineer and SOC Engineer roles in Toronto — love the focus on AWS, AppSec, and detection engineering. The emphasis on open-source contributions and learning culture is also a great signal for high-caliber candidates. I run a recruiting firm focused on security and infra hiring, and I’ve worked with teams placing engineers who’ve built detection pipelines, hardened AWS environments, and deployed tools like CrowdStrike across distributed orgs. If you're open to it, I’d be happy to share a few vetted candidates based in Canada or open to relocating to Toronto. Let me know if that’d be helpful — always happy to support.
Can I get support on this?
Hi Travis, thank you for reaching out! I recommend using the "Generate my Prompt" feature to help refine your prompt. If you find that the prompt itself seems solid but the results are still inconsistent or unreliable, it may be due to the AI model being used. Since you’re currently using the GPT-4.1 nano model, which is the most basic option from OpenAI, you might get better results by switching to a more advanced model. Could you give that a try and let me know how it goes? Cheers -Smit
its still hallucinating pretty bad... im just going to personalize the copy in smartlead for now.
Hey Travis, thanks for reaching out. Got it. I’m not a big fan of letting AI write full emails, since it can hallucinate or miss the mark. I prefer using a set template and then layering in personalized AI-generated snippets for each lead. It keeps things accurate, consistent, and much more controllable. Here’s a quick intro to AI snippets if you haven’t seen it yet: https://www.clay.com/university/lesson/intro-to-ai-snippets-copywriting-automated-outbound Also, when you're thinking about outreach, it really helps to focus on delivering actual value—not just a pitch. Here are two solid examples that do just that: 1. Texada Subject: Your Liebherr LTM 11200-9.1 idle? Dug through some permits and noticed your crane doesn’t seem to be in use (no TX DOT permit found). Meanwhile, Maxwell Construction just filed a permit for a project that needs a crane like yours. That could be a $45k–$180k rental opportunity. Want an intro? 2. On The Stage – Theater Outreach Hi Sarah, Saw your upcoming show, "The Music Man." Found 3 nearby groups who’d likely be interested: ** Riverdale Community Choir (53 members) ** Downtown History Club (78 members) ** Senior Center Music Appreciation Group (41 members) ** All within 5 miles, and here are the contact emails for their leaders.
This kind of outreach creates real value and makes your message worth responding to.
