This 23-Year-Old Thiel Fellow Says the Office Suite Is Broken

The moment of clarity came not in a research lab, but in front of a blank slide deck.

Joseph Semrai was watching his screen fill with widgets, toolbars, and formatting prompts piled on top of each other for decades, none of them truly helping him think. The workflow felt mechanical. He was toggling between apps, searching for the right file, dragging in data from another tab, trying to make sense of information that didn’t want to be organized.

That’s when he realized that productivity software wasn’t designed for thinking. It was designed for typing.

“We have a bunch of disparate applications that aren’t necessarily built keeping the power of [AI models] in mind,” Semrai said in an exclusive interview with AIM Research. “We want to take advantage of the fact that [models] can understand large context windows and use multiple applications at the same time to get the best result.”

That frustration and insight became Context, the AI-native productivity suite he founded in 2024. It’s not just an app. It’s an attempt to rethink the architecture of modern work.

And this week, the market responded. Context announced $11 million in seed funding from Lux Capital, Qualcomm Ventures, and General Catalyst, at a $70 million valuation. With more than 60,000 professionals on its waitlist and a vision to streamline enterprise workflows into a single intelligent interface, the company is quietly laying the groundwork for what may be the most ambitious reimagining of the office suite since the invention of Microsoft Word.

An Operating System for Thought

If you open Context today, you won’t find tabs or toolbars. You’ll find a chat box.

It’s not minimalism for design’s sake. The chat-centric interface acts as the front-end to Context’s proprietary Context Engine, a system Semrai describes as a “reasoning layer” that doesn’t just pull in information but understands what to do with it.

Users can drop in raw documents, financial data, PDFs, spreadsheets, and more and then ask Context to turn them into a board memo, a financial report, or a five-slide summary with charts and takeaways. There’s no need to switch apps, reformat content, or re-upload files. It’s like having an analyst, researcher, designer, and assistant all embedded in the software.

But behind the interface is a philosophical divide. Where incumbents are busy bolting generative AI onto legacy platforms, Semrai’s team is focused on building AI-native infrastructure. “Retrofitting AI into these complex, monolithic systems often yields piecemeal tools that fail to anticipate user intent or handle large contexts,” he said. “Legacy platforms are built on decades-old codebases and business models that monetize incremental feature add-ons and enterprise licensing.”

Context doesn’t play that game. Its AI isn’t there to help you use software but to help you think.

The First Roles to Go

That shift has consequences. When asked what jobs Context might make obsolete, Semrai doesn’t hesitate.

“The first roles to be transformed will be those centered on repetitive content creation: manual slide deck builders, data wranglers who craft basic spreadsheets, and report-formatters,” he said.

But he bristles at the idea that this is about job loss. “We see this not as displacing talent but as allowing knowledge workers to focus on higher-value, strategic activities: strategy formulation, creative problem-solving, and client engagement, while Context handles the busy work.”

In other words, Context is less of a threat to your job and more of a threat to your routine.

Not Another Co-Pilot

In an era when every tool is racing to add a “co-pilot,” Context resists the label. It doesn’t just assist. It drafts, revises, analyzes, and visualizes and it learns what kind of output you’re aiming for.

“Our philosophy is that Context provides near-final drafts in seconds, but the human remains firmly in the loop,” said Semrai. “We augment by shouldering the heavy lifting assembling data, drafting narratives, visualizing charts while users guide direction, inject judgment, and refine nuances.”

And for those worried that such tools might deskill an entire generation of professionals, Semrai makes a counterargument: exposure is education.

“By exposing users to AI-generated best practices in real time, Context becomes a teaching tool,” he said. “Folks can dissect how a professional proposal is structured, study data visualizations, or reverse-engineer formulas. In effect, we accelerate the learning curve rather than short-circuit it.”

Infinite Context, No Noise

One of Context’s most technical feats is its ability to process over 50 million tokens far beyond most commercial tools. But scale without precision can quickly collapse into chaos.

So Context’s core challenge isn’t how much it can see. It’s how well it can focus.

“Our proprietary Context Engine doesn’t just ingest vast text,” Semrai explained. “It intelligently filters, prioritizes, and synthesizes the most relevant signals from both proprietary and public data sources.”

In practice, Context leaves the boundary between automation and human input in the hands of the user. Some will choose to rely heavily on automation to generate near-final outputs quickly, while others may prefer to stay more involved, iterating and refining every step. The platform is built to support both. 

Security as a Feature, Not a Friction

The company’s ambitions also extend into high-trust environments, including government and enterprise. While many startups treat compliance as a bolt-on, Context is baking it in.

“They’re two sides of the same coin,” Semrai said when asked whether security and user delight are at odds. “Industry-leading security and an incredible UX don’t have to compete.”

Context supports on-premise deployment, and it’s building toward enterprise-grade certifications with what Semrai calls “zero-friction security protocols.” “When security is invisible,” he added, “users can focus on productivity, not policy.”

Why It’s Not Just an Agent

With so many AI startups rushing to develop autonomous agents, Context’s decision to build an integrated suite first seems almost contrarian. But Semrai has a clear rationale.

“Agents excel at point solutions, but enterprise workflows span documents, data, presentations, analytics, and search,” he said. “A unified suite ensures context continuity and cross-product integration no need to switch agents or re-upload assets.”

That’s why Context is starting with a full suite rather than diving headfirst into autonomous agents. The team believes that agents are only as effective as the foundation they operate on. Without structured integration across documents, data, presentations, and analytics, agents risk becoming isolated silos powerful, but disconnected.

Context’s approach places agentic capabilities on top of a unified, AI-native foundation. As the underlying models become more capable and reliable, the platform will gradually increase automation but always in service of coherence, not fragmentation.

Why Investors Took the Bet

Context’s vision is undeniably ambitious. But what convinced investors this wasn’t just another AI wrapper?

“Investors recognized that Context represents a fundamental rethinking of productivity software, not merely an AI veneer over legacy tools,” Semrai said. “From our ground-up, AI-native architecture and proprietary Context Engine, every layer was built for scale, performance, and compliance.”

The company’s partnership with Qualcomm also hints at its long-term plans on edge computing, privacy-first inference, and speed optimized for local devices.

For Joseph Semrai, the future of productivity is focused. He believes the true promise of AI lies in clearing away the friction of modern workflows, not adding new layers to them.

“We don’t want people spending their best hours formatting slides or chasing data across systems,” he told AIM Research. “We want to give them back the clarity to think, to decide, to create.”