Founder: AI Copilots Will Replace Office Software

A Bold Prediction Sparks Industry Debate

Karan Mehta, founder and CEO of AI productivity startup Agentic Labs, made waves this week with a provocative keynote at the Future of Work Summit in San Francisco. His central thesis: traditional productivity software — spreadsheets, slide decks, word processors, and project management tools — will be fully replaced by AI copilots within five years.

"Every spreadsheet is just a question someone is trying to answer. Every slide deck is a story someone is trying to tell. Every project board is a plan someone is trying to execute," Mehta told the audience of 2,000 attendees. "AI copilots answer the question directly, tell the story better, and execute the plan autonomously. The middle layer — the software — becomes unnecessary."

"We don't need better spreadsheets. We need to stop making spreadsheets. The AI should go from question to answer without forcing humans to organize data into rows and columns first." — Karan Mehta, CEO of Agentic Labs

The Argument for Replacement

Mehta's argument rests on several key observations about how knowledge workers actually use productivity software:

• 90% of spreadsheet time is data wrangling: Studies show that analysts spend the vast majority of their time cleaning, formatting, and organizing data rather than deriving insights. AI copilots can go directly from raw data to insight.
• Presentations are a compression problem: Slide decks exist to compress complex information into a consumable format. AI can generate tailored summaries, briefings, and narratives for any audience in seconds.
• Project management is coordination overhead: Tools like Jira, Asana, and Monday.com exist because humans need structured systems to coordinate. AI agents can handle task assignment, progress tracking, and deadline management autonomously.
• Document creation is increasingly templated: Most business documents — contracts, proposals, reports — follow predictable patterns that AI can generate with minimal human input.

To support his claims, Mehta pointed to internal data from Agentic Labs' beta users. Companies using their AI copilot reported a 73% reduction in time spent in spreadsheets and a 58% reduction in meetings about project status updates.

Pushback from Industry Leaders

The keynote drew swift and pointed responses from leaders at established productivity companies. Microsoft's VP of Product for Microsoft 365 posted a lengthy response arguing that copilots and traditional software are complementary, not competing.

"Saying AI copilots will replace productivity software is like saying autopilot will replace the cockpit. You still need the instruments, the controls, the visibility. AI makes the pilot more effective — it doesn't eliminate the plane." — Microsoft VP of Product

Google Workspace's product lead echoed similar sentiments, pointing out that AI features integrated into Google Docs, Sheets, and Slides have increased usage of those products, not decreased it. "When we added Gemini to Sheets, daily active users went up 18%. People want smarter tools, not fewer tools," she said.

Several analysts also questioned Mehta's timeline, noting that enterprise adoption of AI tools remains hampered by:

• Data security and compliance concerns, especially in regulated industries
• Resistance from workers who have spent decades mastering existing tools
• Integration complexity with legacy enterprise systems
• Hallucination risks that make fully autonomous AI copilots dangerous for financial and legal work

The Emerging Middle Ground

Many observers see the likely outcome falling between the extremes. Rather than AI copilots completely replacing productivity software, the tools themselves will become increasingly AI-native, blurring the line between "software" and "copilot."

This evolution is already visible. Microsoft's Copilot can now create entire PowerPoint presentations from a text prompt. Google's Gemini in Sheets can analyze datasets and generate charts without the user writing a single formula. Notion's AI can auto-generate project plans from meeting notes.

The question isn't whether AI will transform productivity software — it clearly will. The question is whether the transformation looks more like a replacement (Mehta's thesis) or an evolution (Microsoft and Google's thesis). For professionals navigating this shift, staying current with AI-enhanced tools is becoming a career necessity. Platforms like InterviewAlly help candidates prepare for interviews at companies building these next-generation productivity tools.

The Startup Landscape Betting on Replacement

Mehta isn't alone in his conviction. A growing cohort of startups is explicitly building AI-first alternatives to traditional productivity software:

• Agentic Labs: $85M Series B, building an AI copilot that replaces spreadsheets and BI dashboards for business analysts.
• Tome: $100M+ raised, positioning itself as the AI-native replacement for PowerPoint.
• Granola: $30M raised, building AI meeting notes that aim to replace the need for shared documents and project trackers.
• Dust.tt: $55M raised, building AI assistants that replace internal knowledge bases and wikis.
• Runway Financial: $80M raised, AI-powered financial modeling positioned against Excel.

Collectively, these startups have raised over $2 billion in funding, a signal that investors, at least, believe the replacement thesis has merit.

What This Means for Knowledge Workers

Regardless of which thesis proves correct, the implications for knowledge workers are significant. Proficiency in traditional office software — once a foundational resume requirement — is becoming less differentiating. Meanwhile, the ability to effectively prompt, direct, and collaborate with AI copilots is emerging as a critical skill.

"The most valuable employees in 2028 won't be the ones who know every Excel shortcut. They'll be the ones who can tell an AI what they need and critically evaluate what the AI gives them back." — Harvard Business School professor

A recent LinkedIn Workforce Report found that job postings mentioning "AI tools proficiency" or "AI copilot experience" have increased 340% since January 2025. In contrast, postings specifically requiring "advanced Excel" skills have declined 15% over the same period — a small but meaningful shift that may accelerate as AI copilot capabilities improve.

The debate is far from settled, but one thing is clear: the way knowledge workers interact with productivity software is undergoing its most fundamental transformation since the personal computer revolution. Whether that transformation is led by incumbents or startups, every professional who works with information should be paying attention.