Inside Slack’s Search, Learning and Intelligence team

Head of SLI Noah Weiss on using data to focus our attention in everyday work

A year ago, Noah Weiss started up Slack’s Search, Learning, and Intelligence (SLI) team and New York office, interviewing candidates and working from coffee shops all over the city. Now he leads a team of 12 people applying search and machine learning to build an intelligent layer on top of Slack that gets smarter and makes companies more productive the more they use it.

These features are designed to help customers cut down on basic time-consuming work tasks — like searching for files, hunting for the Wi-Fi password, and catching up on important discussions after hours of meetings — freeing them to do more challenging and creative work.

One year into the SLI team’s work, we’re catching up with Weiss to learn about what the team’s been building, the challenges they’ve confronted, and their goal of giving every knowledge worker a personalized Chief of Staff.

Slack: When you first took this job, you explained in a Medium post that the mandate of Slack’s SLI group will be to “focus on building features that make Slack more effective the bigger a company is and the more it uses Slack.” What does that mean?


Noah Weiss
Noah Weiss, Slack’s head of Search, Learning, and Intelligence


Every month, even a 1,000-person company will probably send millions of messages. Those are now searchable and become knowledge that can be used to help others. That can have a really big impact.


Weiss: One of the biggest issues employees in larger organizations face is feeling like there’s a ton of information to keep up with at work, and it’s overwhelming. While Slack helps with these issues, there’s also an opportunity for us to add an intelligent layer that automatically directs a person’s attention to the information that’s most important to them based on the people, topics, and teams they care most about.

People are using Slack to communicate with coworkers and teams, they’re uploading and sharing files, and they’re using it to pull data from external apps and services. Over time, Slack becomes an organization’s connective tissue and knowledge repository.

Every month, even a 1,000-person company will probably send millions of messages. Those are now searchable and become knowledge that can be used to help others. That can have a really big impact.

Slack: What kinds of things have the SLI team built so far?

Weiss: We’ve made features like Slackbot channel suggestions that recommend channels you should join, or leave, based on what’s most relevant to you. Let’s say you join a new channel, like #developer-team, you might have Slackbot say, “Hey, you might also be interested in #core-product or #data-analytics”, so it’s helping you find your way to the best sources of information.

We also power the “scientific sorting” behind All Unreads, which collects all your unread messages and then prioritizes the order of the messages based on your own past reading behavior.

More recently we’ve been experimenting with search ranking and key signals to improve the quality of Slack’s search results. This year we made our first foray into building universal search sections that present both personalized and recent search results in a single view.


Slack: What have been your biggest challenges?

Weiss: The fundamental thing we’re figuring out is: “How do we want to do data-driven product development work at Slack?” Both in terms of identifying what data we have to power intelligent features and evaluating whether these efforts are making the Slack experience markedly better. Because that question doesn’t just impact us, it affects other teams we work closely with like infrastructure and core product teams.

We’re also solving for a vast problem: information overload. You could potentially work on ideas for years in a lab environment before shipping anything to actual customers. But, we know we need qualitative feedback and real-world usage to know if we’re on the right track. It’s going to require an iterative process where we’re constantly course correcting.

Slack: So, what do the next couple of years of Search, Learning and Intelligence work look like at Slack?

Weiss: When I think of the concept of this intelligent layer, I think a lot about a CEO’s or senior executive’s amazing Chief of Staff. The point of that person’s job isn’t to do all these mundane tasks, like calendar scheduling, it’s actually to help the executive make sure that they’re spending their time on whatever is most important and to have a really good pulse on what that person feels is urgent and interesting.

What we want to do is have Slack, over time, feel like a Chief of Staff for everybody using it, one that can reduce all the time that you waste trying just to keep up with the communication flowing through your organization.

It’ll feel like a real Oprah moment when we get there: YOU get a Chief of Staff! And YOU get a Chief of Staff! And YOU get a Chief of Staff!

It’s all very exciting.

And if you’re keen to join in that excitement (and you happen to have the right skills), by all means check out our current openings on the SLI team here.

Slack is the collaboration hub, where the right people are always in the loop and key information is always at their fingertips. Teamwork in Slack happens in channels — searchable conversations that keep work organized and teams better connected.