7:42pm Friday. Julia notices CPA on the Meta retargeting campaign jumped overnight in the Looker dashboard. She switches to a ChatGPT tab. Pastes the numbers in. Asks what might be going on. Gets an answer. Switches back to Slack. Pastes the answer for the team. Tries to remember what the AI said about why this happened. Switches back to ChatGPT. Asks a follow-up. Pastes that into Slack too.

Half the conversation is in Slack. Half is in a chat tab nobody else can see. The team has fragments. The AI has fragments. Nobody has the full picture.

This is what it looks like when AI lives somewhere different from where the team works.

Marketing teams live in Slack. AI doesn't.

Most marketing AI today ships as a chat window. A separate tab. A sidebar. A canvas. Wherever it lives, it's somewhere else — somewhere you have to remember to open, copy context into, copy answers out of, and re-engage every time you switch tasks.

That's not how marketing teams actually work.

The work happens in Slack. The campaign owner is in #paid-perf. The exec asks pacing questions in #leadership. Brand-safety incidents land in #marketing-incidents. The Friday-night anomaly conversation happens in a thread, with three people piling in. None of that is in your AI's chat window.

A teammate doesn't make you switch tabs to get value out of them. They show up where the work is.

The tab-switching tax

Every time AI lives somewhere different from work, you pay a tax. It's not theoretical — it's measurable, and it shows up in three places.

Lost context. You can paste the data into the chat, but you can't paste in the team's last week of Slack threads, the brand exec's running concerns, the test you launched on Tuesday. The AI is starved of the context that's already sitting two tabs over.

Lost answers. The AI replies, you paraphrase it back into Slack for the team. The team riffs on the paraphrase. By Monday, what the AI actually said is buried in your private chat history. Nobody else can find it. Nobody else can build on it.

Lost momentum. Half the questions don't get asked at all. You see something off in a dashboard at 7:42pm, you think "I should ask the AI about this," but the friction of opening a tab, framing a question, pasting context — it's just enough that you skip it. The questions die.

Bamboo in Slack

Bamboo now lives in Slack. Same agents, same context, same routines — but operating where your team is, not in a separate window.

What that unlocks today:

@boo mentions in any channel. Tag Boo in a thread, ask anything you'd ask in a chat window. The reply lands in the same thread, where your team can see it, react to it, and build on it. The conversation stays one conversation.

Routines post outputs straight to channels. Your daily Radar scan posts its anomaly summary to #paid-perf at 7am. Your weekly performance report drops in #leadership Monday morning. Your pacing alert pings #ops if a campaign drifts off track. The team sees what Boo saw, automatically.

Per-channel agent assignment. Multi-brand teams can scope which agents talk in which channels. Your Nike agent posts to #nike-perf. Your Adidas agent posts to #adidas-perf. Each one carries its own brand context. No cross-talk, no confusion.

DM alerts for specific owners. When something matters to one person — their LinkedIn campaign is pacing 30% over budget, their landing page just broke — Boo DMs them directly. The alert finds the owner instead of waiting for them to find the alert.

Shared visibility. Every action Boo takes is visible to the team via list_history and get_history. The thing the AI did on Tuesday isn't trapped in one person's chat — it's part of the team's record.

A day with Boo in Slack

What this looks like when it's running:

7:00am. Boo's daily Radar routine runs across every connected ad account. Detects that Meta retargeting CPA jumped 38% overnight. Posts a summary in #paid-perf with the campaign, the spike, the likely cause (audience expansion on Tuesday), and a recommended action.

8:14am. The campaign owner opens Slack, sees the post, replies in-thread: "@boo, what was the original audience?" Boo responds in the same thread with the saved audience definition from before Tuesday's change.

9:32am. Pacing alert lands in DM for Lisa — her LinkedIn brand campaign is pacing 30% over the daily target. Boo includes the projected end-of-month spend at current pace.

10:45am. PM in #leadership mentions Boo: "@boo, how did Q1 retargeting compare to Q4 across all platforms?" Boo runs the query, posts the answer in-thread with the chart inline.

1:20pm. The team discusses pausing the audience expansion in #paid-perf. Lead replies "@boo go ahead and pause Prospecting_Broad_US." Boo executes, confirms in-thread, logs the action.

5:00pm. Friday's weekly performance report posts to #leadership. Spend vs. budget. CPA vs. target. Notable changes. The exec sees it land before she logs off, reads it on her phone.

No tab switching. No copy-pasting. No private chat sessions where the AI's value evaporates the moment you close the laptop.

The team works through one shared system, not ten separate chats.

Setup

Connect Slack from your Bamboo workspace. Pick which channels each agent posts to in the Setup tab. Routines you've already configured automatically gain a Slack output sink. New routines have it as an option from day one.

Existing connectors keep working — @boo in Slack uses the same Bamboo MCP tools, the same data sandbox, the same shared knowledge layer that powers Boo in Claude, ChatGPT, and Cursor. The agent doesn't change. The surface does.

Get started

Stop pasting between tabs. Bring your agents to the room.

Try Bamboo free

Bamboo in Slack: Manage your marketing agents as team members