About Notion
A connected workspace where teams build and share their own tools
Notion started as a place to write, organize, and collaborate in one workspace. Over time, teams built on it with docs and databases to run day-to-day work. Today, Notion has evolved to include AI capabilities, where teams and agents work together across tools like Slack.
The challenge
The cost of answering the same questions became impossible to ignore
As Notion grew into a global company with offices in San Francisco, New York, and abroad, its communication systems couldn’t keep pace. The Workplace team — responsible for managing offices, facilities, and the day-to-day employee experience — felt that strain first. Requests that once arrived as a manageable trickle turned into a constant stream.
“When we were smaller, it worked for employees to DM me directly,” said Workplace Portfolio and Projects Partner Tiffany Chan. “As our team grew, it became chaotic.”
Much of the information employees were asking for already existed. Office policies, floor plans, and procedures were documented in Notion, but searching a wiki took more effort than asking a question in Slack. The result was siloed, duplicated work and missed context when people were out of office or offline.
Engineering teams faced a parallel problem. Alerts arrived in Slack, but investigating each one meant manually pulling together clues from multiple systems. “An engineer might have to spend 20 minutes investigating why an alert fired,” said Software Engineer Sean Keenan. Those interruptions added up quickly and broke focus across the day.
“How we operated wasn’t working anymore. Slack freed us up to focus on the things we actually love, like creating magical moments that matter for our employees.”
How Notion works better with Slack
With Slack, Notion turned conversation into an active system of record
Notion has always been the place where teams keep their plans, documents, decisions, and shared knowledge. Slack is where discussions, debates, and decisions happen in real time. Rather than treating those as separate worlds, Notion connected them.
To bridge the two surfaces, Notion brought Notion AI directly into Slack. Teams can @mention Notion in a channel to ask questions, summarize a thread, or pull information from Notion databases — turning live discussion into documentation, plans, or tasks without leaving Slack.
“The thing that’s really valuable about having Slack integrated with Notion is that these are the two places where I do all my work,” said Keenan. “I’m able to search all relevant context.”
Behind the scenes, Notion AI draws from both places. Slack’s search capabilities allow Notion to access relevant messages and threads, while Notion provides the structured, long-term record of work. The result is fewer manual handoffs. Instead of ending with “we should do this,” conversations feed directly into the systems where work is written down, updated, and carried forward.
“I use Slack every day. All the conversations at Notion and within my teams are in Slack.”
A workplace that improves with every question
To further automate work where conversations happen, Notion integrated its own Custom Agents directly into the Slack. These agents pull shared context from Slack and Notion, respond directly in Slack channels, and improve over time through team feedback.
Take Smilers, a Custom Agent that lives in Notion’s #environment-ask Slack channel. What began as a simple intake form is now a digital teammate for the Workplace team that answers employee questions directly in Slack using information stored in Notion.
What makes Smilers different is what happens when it gets something wrong. “We call this a self-healing Wiki,” said Marina Camim, who leads Notion’s Custom Agents. “People can give feedback in the Slack thread or with emoji, and Smilers updates the Wiki and replies back in Slack.”
That seamless Slack feedback loop changed how the Workplace team spends its time. “Smilers saves us at least two to three hours a day,” said Chan. Since launch, Smilers has handled more than 2,500 workplace requests globally, freeing the team from constant manual follow-up and saving 10–20 hours a week.
The engineering team is also taking advantage of this Notion-Slack synergy: Another Custom Agent uses tools like Honeycomb and Splunk in Slack to monitor alerts and provide a first pass analysis before an engineer steps in. “This Custom Agent in Slack reliably comes back with preliminary analysis on what happened, easily saving me more than two hours a day,” said Keenan.
Where human and agents deliver powerful work
Teams build and refine Notion’s Custom Agents together in Slack, shaping how they behave as they’re used. “Having multiple people collaborate in Slack on each agent is really unique,” said Camim. Feedback, corrections, and improvements all happen in this shared environment, not a separate tool. “Understanding how end users actually use Custom Agents is so impactful,” said Keenan. “And that happens in Slack.”
Under the hood, Slack APIs make it possible for the agents to feel like part of the team, both internally at Notion and for Notion’s own users. “When our customers use Custom Agents inside Slack, they expect it to respond like a person,” Keenan said. “And they can do just that with the typing indicators and Slack context.”
That sense of humanity shows up in lighter ways, too. In one of the team’s favorite Slack channels, colleagues share the most surprising or amusing agent responses. Those reactions become another form of feedback, helping teams understand how agents land with real people.
Small details like this matter as Notion continues to scale. “We use Slack all day, every day,” said Chan. “As we’ve grown globally, it’s how we’re able to work quickly and stay aligned across time zones.”
“The thing that makes developing for Slack so impactful is that you can build such a great experience with so few primitives.”
What’s next
Building toward a future where agents coordinate work together in Slack
Inside Notion, Custom Agents are still evolving. The next phase is orchestration: moving beyond individual agents to systems where agents can trigger one another and carry work across longer, more complex sequences of tasks. “We’re just at the beginning,” said Camim. “Having agents work together efficiently in Slack is something we think about a lot.”
As that capability grows, Slack will remain a key interface through which people interact with Notion’s AI. “Slack is where our customers work, and we don’t want to take them out of their flow,” said Keenan. By keeping AI anchored in Slack, Notion ensures agents stay visible, usable, and easy to improve — so questions, feedback, and decisions continue to surface where work already happens.
