Dex Wants to Be the AI Coworker That Actually Lives Where You Work

The Macro: Your Browser Is Already Your Operating System

Here is something people in enterprise software don’t want to admit: most knowledge workers spend their entire day in Chrome. Not in native apps. Not in carefully designed desktop software. In browser tabs. Dozens of them. Salesforce in one tab, Google Sheets in another, Slack in a third, Notion in a fourth, LinkedIn in a fifth. The browser became the OS without anyone planning it.

This has created a specific kind of productivity problem. Your work context is fragmented across tabs, and every time you switch between them, you lose a little bit of momentum. The meeting notes are in Notion but the follow-up action is in Salesforce and the relevant email is in Gmail. You are the integration layer. You are the glue holding all of it together with copy-paste and muscle memory.

The first wave of AI productivity tools tried to solve this by building standalone apps. New surfaces, new dashboards, new places to go. The irony was not lost on anyone: tools designed to reduce complexity were adding another tab to the pile.

The second wave, which is happening now, takes a different approach. Instead of making you come to the AI, bring the AI to where you already are. Live inside the browser. Watch what you do. Learn the patterns. Act across apps without making you leave the page you are on.

That is a meaningfully different product philosophy, and it is the one Dex is betting on.

The Micro: Harvard Dropouts Building Your Browser’s Brain

Dex is a Chrome extension that functions as a context-aware AI assistant embedded directly in your browser. It connects to fourteen-plus apps, including Salesforce, Google Sheets, Slack, Notion, Linear, and LinkedIn, and learns from your interactions to anticipate what you need next.

The core pitch is that Dex remembers your preferences and context across sessions. If you spent Tuesday building a lead list by scraping LinkedIn profiles into a spreadsheet, Dex noticed. Next time you start a similar workflow, it can handle the mechanical parts while you focus on judgment calls. Meeting prep is another use case they emphasize: Dex pulls from your calendar, email history, and LinkedIn connections to brief you before a call.

What separates this from the fifteen other “AI assistant” Chrome extensions is the multimodal workspace angle. Dex is not just answering questions or generating text. It is watching your workflow across applications and building a model of how you work. That is closer to what Cursor does for developers than what ChatGPT does for everyone else. The tagline “Cursor for knowledge work” is ambitious, but at least it communicates the intent clearly.

Regina Lin and Kevin Gu founded the company. Regina studied mathematics with a concurrent master’s in computational science at Harvard before dropping out to build this. She previously co-founded Contour AI and worked in tech investment. Kevin’s background is quant and ML research, with stints at Jump Trading, IBM Research, QuantCo, and Meta. They are a two-person team out of San Francisco, part of YC’s Winter 2025 batch.

The privacy angle is worth mentioning. Dex emphasizes granular permission controls and claims user data stays private. For a product that literally watches everything you do in the browser, that is not a nice-to-have. It is table stakes. If enterprise buyers don’t trust the data handling, nothing else matters.

Pricing includes a free tier with Pro, Business, and Enterprise plans for advanced features.

The Verdict

I think the browser-native approach is correct. The problem with standalone AI productivity tools is that they add friction to a workflow that already has too much of it. Living inside Chrome, where the work actually happens, removes the biggest adoption barrier.

The risk is that this space is getting crowded fast. Every AI lab and their cousin is building Chrome extensions. The moat here has to be context depth. Can Dex actually learn your patterns well enough to anticipate actions, or does it end up being a slightly smarter autocomplete? The technical challenge of maintaining context across fourteen different apps, each with their own data structures and authentication, is substantial.

Thirty days from now, I want to see daily active usage numbers. Not installs, not signups. How many people open Chrome and actually interact with Dex more than once? Sixty days, show me the cross-app actions. Are users doing multi-step workflows through Dex, or just using it as a search bar? Ninety days, the question is whether enterprise teams are adopting it or whether this stays a power-user tool for individual operators. The Cursor comparison is flattering but the gap between “good Chrome extension” and “indispensable workflow layer” is real and it is wide.