Personalized Cold Email Sequence: Build AI-Driven Outreach with Clay and Smartlead

Start in Clay with a focused list of 50-150 prospects. Use Clay's company and people search to filter by job title, company size, industry, and geography. Resist the urge to go wider than 150 on your first sequence: smaller, tighter lists produce better personalization and cleaner signal on what messaging actually works. For each prospect, you need at minimum their name, company name, company domain, and LinkedIn URL. Add the company description column from Clay's enrichment waterfall. This becomes the raw material for your AI personalization layer in the next step.

This is the step that separates good cold email from spam. In Clay, add an AI column powered by Claude or GPT-4. Write a prompt that takes the enriched data (company description, job title, LinkedIn headline, recent funding or hiring signals) and outputs a personalized first line for your email. A strong prompt structure: "Based on the following company information, write a single conversational sentence that references something specific about their work or growth stage. Do not mention their product name generically. Be specific and direct." Add a second AI column for a "personalization angle": one reason why this person, at this company, would care about your offer right now. Use both columns in your email template.

Export your enriched Clay table as a CSV, then import into Smartlead as a new lead list. Before you create the sequence, check your sending infrastructure: your email accounts should be warmed up for at least 3-4 weeks before sending cold outreach. In Smartlead, configure your sending limits at 20-30 emails per day per account during the first two weeks of a new sequence. Connect multiple sending accounts if you have them and use Smartlead's rotation feature to distribute volume across accounts. This keeps your daily sending per domain below spam-trigger thresholds and protects deliverability.

Build a 3-email sequence in Smartlead. Email 1 is your opener: personalized first line (from your Clay AI column), a one-sentence description of what you do, and a low-friction call to action. Keep it under 80 words. Email 2 is a value-add follow-up sent 3-4 days after Email 1: share a specific insight, case study result, or relevant piece of content. No pitch. Just evidence of expertise. Email 3 is a break-up email sent 5-7 days after Email 2: short, direct, and genuine. Something like "Not sure if the timing is right, but happy to reconnect whenever it is." Use your Clay personalization angle variable in the subject line or opening of at least Email 1 and Email 2.

Once your sequence is live, track four numbers daily: open rate, reply rate, positive reply rate, and unsubscribe rate. A healthy cold sequence on a warm sending domain should see open rates above 40% and reply rates above 3%. If open rates are low, the problem is deliverability or subject lines. If reply rates are low relative to opens, the problem is your offer or personalization. Do not run a sequence for more than two weeks before reviewing results. After you hit 50+ replies (positive and negative), you have enough signal to decide whether to expand the campaign or rewrite the sequence. Iterate on one variable at a time so you know what actually moved the needle.