Guide · made with Zapier
How do I build an AI agent that does real work for me, without writing code?
Nick built a working research agent in Zapier Copilot from one plain-English prompt: it finds five guests every Monday and drafts the outreach emails.
The short answer
You describe the job in plain English and let Zapier Copilot build it. Nick typed one paragraph asking for an agent that researches potential AI guests, saves their details to a Google Sheet, and drafts invitation emails in Gmail every Monday, and Copilot proposed the system, wrote the workflow instructions, set a weekly Monday 9:00 a.m. trigger, and connected Sheets and Gmail on its own. He wrote no HTML and no Python. The human work that was left was small: connect the accounts, create the Google Sheet himself because the agent could only add rows to an existing sheet, add the column headers, point the Sheets action at the right file, then approve each action during the test run. On the first test it found Melanie Mitchell, Lex Fridman, Andrew Ng, Yann LeCun, and Max Tegmark, filled the rows in real time, and put a finished draft in Gmail.
Build your own agent with Zapier →Built with Zapier, the sponsor of this video.
Step by step
Open the Zapier Copilot builder and connect your apps first ▶1:15
Nick opened Zapier with Copilot from the link in the video description, which loads the agent builder with the prompt already in it. His Gmail and Google Drive were already connected to Zapier, so the agent could reach Gmail and Google Sheets. If yours are not connected, do that first. Linking an app is just logging into it through the Zapier interface and granting access.
Write the job as one plain paragraph ▶1:32
Do not try to write instructions like code. Nick typed one block of normal English: create an agent that researches potential guest experts on artificial intelligence, people with large platforms, a book in the space, or podcast appearances about AI. He named the exact spreadsheet columns he wanted, first name, last name, date discovered, follower count, social accounts, contact email, a summary of accomplishments and why they fit the show. Then the cadence and the output: find five potential guests each Monday and draft an email to each one, saved under drafts in Gmail.
Read the proposed system before you continue ▶2:38
Copilot came back with a proposed system, a capability list, the apps it would integrate, and a section called assumptions made. Read the assumptions. That is where it tells you what it decided on its own, like using web search to find experts and running the automation weekly. Nick realized while reading it that he had left something out, so he added a line asking it to also look for people with books about to come out in the space, and it folded that into the capabilities.
Continue to Agent Builder and let it fill in the details ▶4:06
Clicking continue with agent builder hands the plan to the agent system, which writes the workflow instructions and builds the steps. It set a Schedule by Zapier weekly trigger, and when it needed a time of day it asked itself the question and picked Monday at 9:00 a.m. without being told. Nick's read on it: Copilot and the Zapier agent system are talking to each other and working out the details with no human in between.
Keep the last step a draft, never a send ▶6:48
The agent saves finished emails to Gmail drafts instead of sending them. That is deliberate. Nick does not plan to ever let an AI send email on his behalf, no matter how well it writes in his voice, because he wants to make edits and keep the human element. The goal is to remove the drudgery, not the judgment.
Create the Google Sheet yourself and add the headers ▶7:39
Nick asked the agent to create a fresh spreadsheet so the data would not land in an existing one. It could not. It can only add rows to spreadsheets that already exist, so it asked him to make the sheet. He created one, named it for the project, and added the column headers it listed. Small piece of manual work, and you have to do it before the run.
Point the Sheets action at the right file when it cannot find yours ▶9:21
The agent finished the Gmail sequence but still could not see the new sheet. It told Nick the core research and outreach functionality was ready and offered to save the agent and fix the sheet afterward. He opened the agent settings, opened the create spreadsheet row action, searched for his sheet by name directly, and selected it. Then he let the agent pick the worksheet value itself.
Preview and test, approving each action as it goes ▶10:51
Hit preview, then test agent. It ran web searches, pulled bios, social accounts, follower counts, and contact details, and asked for approval before each action. Watch for limits: on the free agents plan Nick hit a cap of 10 activities per run, with a paid plan offering up to 40 continuous activities. Check the spreadsheet while it runs to confirm rows are actually landing.
Turn the agent on and name it ▶21:17
A finished agent is not a running agent. Nick had to click the enable button to switch it on, and only then noticed he had never named it, so he named it after the job it does. Your agents list shows everything you have built. He also accepted the agent's own suggestion to add a step that checks the existing sheet before each run so it does not pull duplicate names.
“Think about it as working with an intern that needs to be managed rather than an executive who can come in and just handle everything.”
Nick Standlea, in the video
Key moments
- ▶0:00Why build an agent at all
- ▶1:32The exact prompt Nick typed in
- ▶4:06Continue to Agent Builder
- ▶6:48Drafts only, never auto-send
- ▶7:39The agent cannot create a sheet
- ▶10:22Manually pointing it at the sheet
- ▶10:51First preview and test run
- ▶12:42Free plan caps activities per run
- ▶18:22Checking the finished Gmail draft
- ▶21:17Enabling and naming the agent
Questions people ask
can you build an ai agent without coding
Yes. Nick built a working research and outreach agent in Zapier Copilot by typing a single paragraph of plain English describing the job. No HTML and no Python, and he says outright that he does not know how to write either. Copilot wrote the workflow instructions, set the trigger, and wired up Google Sheets and Gmail.
what can a zapier ai agent actually do
Nick's agent runs every Monday at 9:00 a.m., searches the web for five potential AI expert guests, evaluates them against his criteria of large platforms, books, and podcast appearances, writes their details into a Google Sheet, and drafts a personalized invitation email for each one in Gmail. On the test run it surfaced Melanie Mitchell, Lex Fridman, Andrew Ng, Yann LeCun, and Max Tegmark, with bios, social accounts, follower counts, and contact information. It makes real decisions about who is worth including rather than just moving data around.
how do I write a prompt for an ai agent
Describe the outcome the way you would explain it to a new hire. Nick's prompt named the task, the selection criteria, every spreadsheet column he wanted, how many results per run, how often to run, and where the output should land. Being that specific is what let Copilot propose a complete system on the first pass. When he realized he had forgotten a criterion, he simply added another sentence and it updated the plan.
should I let an ai agent send emails for me
Nick says no, and he built his agent to stop at the draft. His words: he does not plan to ever have AI send emails on his behalf, no matter how well it can write in his voice, because he always wants to make edits and keep the human element. The agent removes the research and the first draft. You still review, personalize, and hit send.
why can't my zapier agent find my google sheet
This happened to Nick on camera. The agent cannot create a new spreadsheet, it can only add rows to one that already exists, and after he created his sheet the agent still could not see it. The fix was manual: open the agent settings, open the Google Sheets create spreadsheet row action, search for the sheet by name directly, and select it. He saved the agent without testing until that was corrected.
is the zapier agents free plan enough to test an agent
It was enough for Nick to build and start testing, but he hit a ceiling mid-run. The message said he had reached the limit of 10 activities per run on the agents free plan, and that a paid agents plan unlocks up to 40 continuous activities per run. His take was that if the agent does what he hopes, the price is worth it.
how much do you have to supervise an ai agent
More than zero, and Nick is direct about the frame. Treat it as an intern who needs managing, not an executive who handles everything. He had to connect accounts, create the sheet, add the headers, fix the Sheets selection, and approve each action during the test, and he still plans to check every guest the agent finds before reaching out. The point is how little intervention it took, not none.
Zapier sponsored this video. Nick chose what to build and what to say.