From Systems of Record to Systems of Context — Omri Bruchim & Tomer Ast, monday.com

Um, hey everyone. We are super excited to be here, thanks for having us. Today we're going to talk about how we shift monday.com from a system of record into a system of context. And honestly, the title tells the whole story in just a single line.

For decades we built software that records what happened: every task, every document, every message, every status update, just put into the record. What we want to talk about today is, like, take it a step farther. We want software that actually understands the connection between them.

So I want to start from a simple question that each one of us asks himself every morning: what should I focus onright now? It sounds almost trivial, but to be honest with yourself, if you ask your AI agent, whether it's Gemini, GPT, or even Claude, if you ever typed this question, you probably got a list of bullets not related to each other.

A list of, like, items dressed up like a confident paragraph. But it's not really connected to what you're working on. Actually, I tested it last week and Claude asked me to go to the gym. I don't know if it's a compliment or not, but this is what he suggested.

And what really makes it so frustrating is, like, your AI assistant has all this data. It has all the boards, the tasks, the emails, the Slack messages, everything that you ever touched if you connect it. But it still can't really answer it.

It has all the data, but it has zero understanding. So this is the real challenge we are facing. This is the heart of the entire talk. The problem was never the missing of data, the retrieval. The problem is, like, the missing understanding.

Those are two totally different things, and almost everyone mixes between them. Understanding is the word that we're going to focus on the entire talk. Not context, not memory, not retrieval. Understanding. So, um, quick about ourselves. My name is Tomer.

My name is Omri. This is Tomer. We both engineer and manager at monday.com, working on exactly the problem that we're going to talk about. A little bit of context about where we're coming from. Monday.com is a global software company.

We build the AI work platform used by hundreds of thousands of teams. And the part that really matters for the talk is that Monday is where the work lives. We help companies to do the work, not just, like, save new records.

Every project, every task, every decision, every meeting, the notes, the action item, everything logs into the system. And our mission has always been to help the teams to achieve their business outcome. Whether you are a salesperson, so we want to help you, like, create an SDR agent that calls to your prospects and helps you sell the product.

And if you are, like, a finance team or marketing, we want to help you with research. So each one of the disciplines, we want to help you to do the job. And together we have, like, four big bets on the AI work platform.

We have, like, Monday Sidekick, we're going to talk about it mostly today. Monday Vibe, if you want to build your own software. Monday Agent, if you want to create your own agent into the platform. And, like, if you want some more deterministic flow, we have, like, Monday Workflows.

But today we're going to focus on Sidekick. Sidekick is your intelligent AI personal assistant that understands your work, thinks, and executes with you. It's like a bird on your shoulder. He knows you, he knows your business. It helps you with your work on every aspect.

He works the way you do, with your tones. It keeps you totally in control. But, like, putting all these bullets and building all these promises is really hard. So let's talk about why it's hard. There are three points for that.

First, like, picture one assistant sitting on top of absolutely everything. You have your Slack messages, you have all the notes for your meetings, absolutely everything. And it's beautiful and overwhelming at the same time, because as it stands, like, there's a wall of records everywhere.

So why is it making it so hard to connect between them? There are three reasons for that. One, it's something that we call the agent gap. Your agent, every agent, is really sharp and doing tasks when he knows what to do.

But he sometimes lost finds them, finds the problem. If I ask my agent, like, please help me draft and reply to, like, the escalation from a customer, he nails it. He knows how to do it. He takes the context.

He likes doing the work. But asking him, what should I focus on first? He guesses, because he doesn't understand what is my priority. Who am I? It doesn't matter if he has a memory or something. He still doesn't know what is the problem.

The second problem is, like,

we have the records, but we don't have the meaning. A log never says what it means. Let's take an analogy from our world. Let's say that you have one line of code in your GitHub. You look at the code.

Maybe there is a comment on top of it. Nice. But you don't really understand why someone wrote this line of code. Today, none of us write code, but somewhere in the past, someone wrote this line of code. And if you really want to know, you can go to git blame and understand from the commit log why someone did it.

But if you want to go farther, you can go to the PR and maybe read the description. And if you really want to go hard, you can go to your Monday board and see which PR is connected to this item and understand that this line of code came because some customer complained about something.

So this is what we are trying to build. And the third point that is a bit challenging as well is, like, it's really hard to build ahead of time, at runtime, the meaning of things. It's really too late.

You simply can't do something like that the moment someone asks the question. So understanding of the context has to be ahead of time. You need to build it much before someone asks the question.

And this is why we are building what we are building. We are building the Monday world model. This is what we call it. The Monday world model helps you understand why this matters, how to help you, who you are, when, and what not to do.

It's the context that follows your work. He understands who you are. And it's simply not a bigger prompt. It's not a longer context window. It's totally different from what we know until now. First, before, like, Tomer is going to talk about how we build it, what it's not.

It's not a retrieval problem. The problem has never been getting the data. We have all the data. We have all the connections to watch. All the providers that we want to get, all the MCPs. The problem is really to understand how it works.

Understand how each one of these entities is connected to each other. So go ahead, Tomer.

Thank you, Omri. Hi everybody.

So what's the data model? We collect thousands of data points on the user. Every item status change, their activity log, messages and meetings, and construct three things the agent can reason over. The first is how the user's work is structured.

There are key entities and relationships and connections between them. What depends on what? How a message in Slack connects to a task. Who's blocking whom? The second is a current snapshot. Live signals over those entities. What's overdue? What's criticallyurgentright now?

Which coworkers you've been actively working with and why? The third is what we can learn about the user over time. There are decisions and outcomes, work patterns and cadences, distilled into a durable profile. So how do we build that data model?

We use two engines running on different time windows and schedules. A slow engine that runs on a long time window and learns the user and their work, and a fast engine that reads what's happeningright now and how it affects the user's work.

One knows you, and the other one knows your day. First, the slow engine. It takes as context user's activity over weeks and minds it for patterns and the type of persona the user is. Their routines, their work rhythm, who they collaborate with, their main goals, and current projects.

Those patterns get distilled into a durable profile, and every time a profile holds, it's reinforced. This engine tries over time to learn who exactly the user is and how they work.

The fast engine is the opposite. It takes as context a short recent window and recomputes a set of live signals over the user's current state. What's overdue? What's suddenlyurgent? Which coworkers you've been pulled in with? This engine tries to understand your day and updates frequently.

This split isn't something we invented. It's present in two totally different fields. In neuroscience, this split is referred to as complementary learning systems. And in data processing architecture, it's referred to as the Lambda Architecture. We apply the same concepts to how we construct the agent's data model.

Our brain uses the same split. Every important experience gets captured instantly by the hippocampus, and over time the neocortex distills those into durable lessons. In data infrastructure, there's the same split. A fast split layer over a recent real-time window and a slow batch layer over the full history that gets recomputed, and the two are merged into a single served view.

Two different fields landed on the same idea, and that's what we're trying to apply to our data model. So how does it all come together? We collect data from everywhere our users work: Monday, Slack, emails, calendar. And we turn it into data structures, signals, and patterns.

Both engines precompute on top of that, offline and ahead of time. And when a user engages with Sidekick, a thin slice of logic is recomputed for recent activity, and the entire context is served to the agent. Sidekick can then decide when and how to traverse and retrieve context from the data model itself, and it's primed to reason on top of it.

And building it this way gives us two behaviors out of the box. The data model, it's resilient. Sources are isolated, so a bad feed can't break the rest. And the thin layer of logic that runs at serve time verifies part of the context against live data, while the rest falls back to the last verified context.

So it degrades gracefully, but it doesn't fail. Second, it actually understands theurgency of facts. It's able to understand when and how it should be proactive and notify the user, and when it should stay silent.

And the crucial part is that it compounds. Every day the data is captured, the layers fill in, and the profile sharpens. And adding a new data source is deliberately cheap and only contributes, so the surface only grows. The more it sees, the more it understands.

And the more it understands, the more you can lean on it. And the data model is unique. It's unique to how you work. We're not pretending this solves everything. The model itself is always trailing the actual live world.

New users have no reliable data to reason from yet, and signals have our own biases built in. So the hardest part is actually telling the important parts from the noise. But this is an architectural design that we can enrich and test and improve over time.

Thank you.

So back to the original question that we had. What should I focus onright now? Remember the question. So now Sidekick can really answer that. Let's see how. So first, like Tomer said, we collect all the data points. We call it breadcrumbs.

For example, the board that I'm working onright now, all the emails from the past month, the transcript from the meeting, all the action items that I got from this meeting, and we process them offline. We see a full picture of your calendar.

This is what helps us to understand the pattern. So if you have, like, biweekly with your VPs, I understand that tomorrow you have another meeting. Even Slack messages from the last few days to understand if you say something in Slack that's relevant that you didn't say in your daily standup with the team.

We take all this data and we process them either fast or slow. The slow builds some kind of profile on you. So you can see it on the top, like Omri Bruchim is an engineering manager. I'm working on two projects.

Sidekick is a note taker. I'm from Tel Aviv. This is how many hours I have per day and, like, how I work every day. And the fast is something that's helped me to understand what is the action item.

I have, like, three commitments that I promised to other people. I have to reply to a VP that sent me an email, and I didn't. So this is only from the past day or few days window. This is something that's relevant more for today.

And this is how Sidekick can answer those questions. So to summarize it, the bottleneck was never the capability where we take all the data from. It was the understanding. How you connect each one of the dots to each other.

The most capable agent in the world, whether it's like Claude and Gemini, it doesn't understand you. He needs to process it beforehand. This is what we're trying to solve at Sidekick. This is the Monday world model. This is what we are building.

If you have any questions, we are out of time. We are here near the stage. Thank you very much for the time. You can follow us on LinkedIn. We post things on the area. Thank you very much.