In What Order Should I Add Tools to My AI Agent?
The build on my session page had only Gmail and Google Calendar. What the page lists for adding everything else, and the rule that comes before all of it.
Set the approval rule first, then add one tool at a time. My session page lists them in this order for after the first build: Exa and Firecrawl, then Google Docs, then Slack. The build the page walks through only had Gmail and Google Calendar.
Your first AI agent works, and now every tool you've heard of looks like the next thing to connect. Adding them all at once is the one move I'd skip. The order comes from the page I published for the October 7, 2026 session.
What did the first build contain?
Gmail and Google Calendar, or a pasted list if you'd rather not connect. That's what the steps on the session page ask for, and the page says to check that only those two are in the connectors list before you go further. On the recording, I told the group to stay in their comfort zone:
From the recording, 27:00. If you feel comfortable setting up an email account and connecting your agent to it today, you can do that.
From the recording, 27:06. If you're not yet comfortable, just continue to follow along.
I named the systems I'd already given an AI agent access to, such as my Google Drive, Sheets, Gmail and a calendar (31:04).
What does the session page list?
The session page lists what to add after today, from your own account, with a reason for each. It says to add them one at a time and doesn't say why one comes before another:
| As listed | What to add | The page's reason | |---|---|---| | Before any of them | The approval setting | So nothing sends or posts without you | | 1 | Exa and Firecrawl | So it can research a person or company before your call | | 2 | Google Docs | So the brief lands in a document | | 3 | Slack | If you run a team |
I talked about Exa and Firecrawl in the session (27:17 and 29:24). I didn't connect Google Docs or Slack on the recording, so the reasons for those two come from the page.
Why does the approval setting come first?
Because the page puts it ahead of every tool on the list, so nothing sends or posts without you. I showed where the permission lives in tool management: under write and delete tools I can give permission or restrict access to sending emails, updating, creating drafts and forwarding emails (44:19).
The session page says an AI agent with an inbox connector and no rule can send a draft you never read. The instructions file on the session page has a written rule, and it says how to add anything new:
From the file on the session page. It does not connect to Slack, a CRM, or any other tool. Once you know it well, add those one at a time from your own account, with the same rule: never send, never post, without me saying so in that turn.
The file's rule is written text. The approval setting is the control in tool management, so the page puts it first.
Why one at a time?
The file says to add them one at a time. In the build I held one back. My first prompt said I'd use Slack, and later I changed my mind:
From the recording, 38:10. In my initial prompt, I said I'm going to use Slack, but now I'm pausing.
Then I told my AI agent to stop asking (38:16), and you can do the same.
What does the documentation add?
The builder in my session was Gumloop. On my relationship to Gumloop, I said this:
From the recording, 33:22. I don't work for Gumloop. I found this tool over the summer through Lenny's podcast, and I started using it and have since made myself an ambassador for them. I've created a relationship with Gumloop. I am not an employee.
More on that is in Do you work for Gumloop? (the answer from the session).
Checked on 2026-11-05 against Gumloop's AI agent documentation. It says the connectors you add define what the AI agent can see and do, and to start with the two or three it actually needs. Checked on 2026-11-05 against Gumloop's Human in the loop page: you choose the approval setting per connector, and the page lists Always allow as the default.
What goes wrong
A tool that can send, added before the rule. The session page says that without a rule, an AI agent with an inbox connector can send a draft you never read.
A connectors screen that doesn't close until you save. I said you have to click the save button just to exit it (31:20).
I didn't show a second tool being added after the first build.
Where this comes from
This article grows from the session I taught live with Women Applying AI on October 7, 2026. The quoted passages from the recording are my own words, taken verbatim. The passage marked as from the file on the session page comes from the files published with that session. The steps, tables and checklists around them were drafted by my team's AI agents from those words and my published pages, and checked against the recording before publishing. Screens change, and platform facts that have a public source carry the dated line above.
Build Your First AI Agent in One Hour: the full session with the clips, the prompts and the transcript.
Related reading: Which Connectors Does an AI Agent Need?, How Do You Control What an AI Agent Can Do in Your Gmail?, Exa, Firecrawl or Apify: Which Should My AI Agent Use?.
Frequently asked questions
What should I add first?
The approval setting. Then, from the session page's order, Exa and Firecrawl.
Should I connect everything at once?
No. The file on the session page says to add them one at a time, and Gumloop's documentation says to start with the two or three the AI agent needs.
When does Slack come in?
Last in the page's order, for people who run a team. In my build I paused on it (38:10).
What did the first build have?
Gmail and Google Calendar, or a pasted list.
Build something that actually runs your workflow.
A focused, free 60-minute live session with Jacqueline. You build alongside her, on your own real task, and leave with an AI agent that is already running.