How I Run My Day With a Team of Grok Bots
Most mornings I open one status roll-up, not five chats. Overnight leads are already in a CSV. Slack and Jira noise is already turned into tickets. That is how I run a day with a team of Grok Bots.
I run several Grok Bots as a small team — and I am still trying new connections and workflows every week. Some bots are stable parts of my day: content, a PA across Slack, Jira, and Gmail, ConnectEdu campaigns, overnight lead lists, and a group roll-up that feels like a scrum. Others are experiments: video review with voiceover and subtitles, messaging connectors I tried and sometimes pulled back from, and deeper product actions inside ConnectEdu’s Admin AI Agent.
The point is not a fixed stack. It is a lab that earns its place when it saves real time.
The rule stays the same: bots prepare and execute inside clear boundaries. I decide what goes live, what gets assigned, and what needs my eyes.
This is a team — and a lab
My day mixes product, marketing, school SaaS ops, and outbound. One mega-bot made context messy and permissions risky. So I split roles like a human team, then kept adding experiments when a new pain showed up.
Stable desks today:
- Content and SEO
- Personal assistant across Slack, Jira, and Gmail
- ConnectEdu Super Admin campaigns
- Lead generation that runs around the clock
- A group bot that gives me scrum-style status
Recent experiments:
- Video review, voiceover, and subtitles for tutorial videos
- Messaging connectors (WhatsApp worked, then I stepped back — more below)
- Deeper product actions inside ConnectEdu’s Admin AI Agent
Same platform. Different jobs. Different connectors. Always under review.
Bot 1 — Content and SEO desk
This one owns the writing and publishing queue.
It holds a topic queue for products like ConnectEdu, writes full drafts, and — only when I approve — turns them into site pages with meta, nav, and a sample hero image, then commits and pushes.
For personal or brand essays on InRay Labs, drafts stay in chat until I say publish. Speed with a gate.
Bot 2 — PA across Slack, Jira, and Gmail
This bot is my operations PA.
Connected to Slack, Jira, and Gmail (and calendar where needed), it:
- Checks updates across those tools
- Creates Jira tickets and assigns them to the right developer
- Watches Slack channels and flags what I actually need to see
- Checks Gmail and calendar and updates me on what matters
I do not want another inbox. I want noise turned into tickets, assignments, and short alerts.
Bot 3 — ConnectEdu Super Admin campaigns
Another bot is wired to ConnectEdu Super Admin.
It can check campaign status, help create new campaigns, and send emails to leads from our lead pipeline. School outreach keeps moving without me living in the admin panel all day. Brand sends still need my judgment.
Bot 4 — Lead generation, 24/7
One workflow generates leads from multiple public platforms.
That bot runs around the clock. While I sleep, it gathers leads, structures them, and produces CSV files for the next morning. It does not replace sales conversation. It removes the grind of building the list.
Bot 5 — The team bot (scrum for agents)
I keep a group where these bots work together. That bot updates me like a scrum in Agile — what moved, what is blocked, what needs me.
Instead of five threads, I get one standup-style picture of the agent team. That is how I manage the day: a multi-agent desk with a daily roll-up.
Experiment: tutorial videos — review, voiceover, subtitles
This is the part that is not “only the core five.”
My designer produced tutorial videos that were not final for me. I asked one of the bots to go through each video, understand what was happening on screen, write the voiceover, prepare subtitles, and download everything into a dedicated folder on my computer.
It did that neatly. That is the experiment mindset: if a repetitive media task eats hours, I try a bot on it before I hire more process.
Experiment: WhatsApp (and why I disconnected it)
I also tried connecting messaging apps — WhatsApp and Telegram.
WhatsApp connected successfully at first. Then I got a warning in WhatsApp that I believe was related to that connection, so I exited and disconnected it. I am not using the WhatsApp connection now.
That is part of the story too. Not every integration stays. Some work. Some create risk. You pull the plug, learn, and try the next channel carefully.
ConnectEdu Admin AI Agent — answers and actions
When I write about ConnectEdu, I keep product claims accurate.
The Admin AI Agent answers from live school data via secure database queries. Staff can ask in any language; answers come in English.
And it is not answers-only anymore. Actions are already implemented — including adding students, adding staff, updating students, and updating staff. The agent is part of real school operations, not a demo that only chats.
How a normal day looks
Morning. Team status. Overnight leads CSV. Slack or Jira items that need a human. Campaign health. Next content item. Sometimes a video or subtitle batch waiting in that folder.
Daytime. PA turns important Slack or mail into Jira work. Content bot drafts or publishes only what I approved. Campaign bot prepares outreach I have cleared. Experiments get a time box — they do not eat the whole day.
Night. Lead gen keeps collecting. I review output when I am back.
What I refuse to hand over
- Final publish for brand or personal posts
- Product truth — no hype beyond what is live
- Anything that sends as me or emails schools without a clear ask
- Fake metrics or invented quotes
- Risky connectors I am not willing to leave running (WhatsApp is the current example)
What actually changed
Less tab-switching between Slack, Jira, Gmail, admin panels, and a blank page.
Parallel work on leads, drafts, tickets, and now media cleanup.
A scrum-style view of the agent team — plus room to keep experimenting.
What did not change: ownership. The bots run plays. I still call the game.
Need help implementing Grok Bot?
If you want help designing or implementing a Grok Bot setup like this — content desks, PA workflows across Slack/Jira/Gmail, campaign ops, lead pipelines, video helpers, or a multi-bot scrum group — contact InRay Labs.
We build practical AI agent workflows for school and SaaS teams, not slideware.
InRay Labs: https://inraylabs.com · hello@inraylabs.com
