RecipeAgents / Guides
Agent recipe guideAI lead research agent
Personalized outreach works, but researching each company takes 15 minutes you don't have. A lead research agent does the first pass: you add a company to a sheet, and it writes a brief and a specific opener back to the row.
This recipe uses Clay or Perplexity for research and Google Sheets as the queue.
The prompt
Research {{company}}. Return: what they sell, ICP, recent news (with links), 2 likely pains our product [X] solves, and a 2-sentence cold opener referencing something specific.Step by step
- Add a company name or URL to a new row in a Google Sheet.
- The agent researches the company's site, news and LinkedIn presence (Clay enrichment or a Perplexity call).
- It writes the brief and the personalized opener back into columns on the same row.
- You review, edit the opener, and send from your own email or sequencer.
Why ask for links
Research models can be confidently wrong. Requiring links for news items lets you check the claim in seconds before you mention it to a prospect. If an item has no link, treat it as unverified and leave it out of your email.
Getting better openers
Fill [X] with a one-line description of your product and its main outcome. Add an example of an opener you liked. Ask for the opener to reference one specific thing from the brief, not a compliment about the company.
Mistakes to avoid
Sending openers unedited at volume; they start to sound alike.
Mentioning personal details about individuals found online. Stick to company-level facts.
Skipping the link check on news items.
Build it in Make, Zapier or n8n
The recipe is tool-agnostic. Every step maps to a standard building block: a trigger from the app where the data starts, an AI step that sends text to a model with the prompt, a parser that reads the JSON or text that comes back, and actions in the destination apps. Zapier is the quickest to get running for linear flows. Make makes branching and loops easier to see. n8n can be self-hosted if you want to keep data on your own infrastructure. Pick the one your team already pays for.
Test it before you trust it
Run the recipe on ten to twenty real examples before you switch it on for everything. Put the inputs and outputs side by side in a sheet and mark each one right or wrong. Where it's wrong, add one specific line to the prompt that addresses the pattern, then run the same examples again. Keep a human approval step on any action that is visible to customers or touches money until the error rate is one you'd accept from a new hire.
Watch running costs as well. Each AI step is billed by the model provider and each task or operation by the automation tool. Filtering inputs before the AI step, for example skipping newsletters or tiny files, is the easiest way to keep both down. Check current pricing on each vendor's site, since it changes often.
Before you use it
Every recipe on RecipeAgents follows the same shape: a trigger, the tools, the steps, and the exact prompt. Start with the human-approval version. Once you have watched a few dozen runs and trust the output, decide which steps can run on their own. Model and tool names change often, so check current pricing and features in each tool before you build.
Frequently asked questions
- Clay or Perplexity?
- Clay is built for enrichment at scale inside a table. Perplexity is a simple way to get cited research for a smaller list. The recipe works with either.
- Can this run automatically for every new row?
- Yes, with a Sheet trigger in Zapier or Make. Keep the send step manual.
- Does it work for B2C leads?
- It's designed for companies. Researching private individuals raises privacy issues we don't recommend.