A few years ago, enterprise AI adoption felt like pulling teeth. Founders struggled to show teams how AI could actually do their jobs, not just assist. But according to Akshay Nathan from OpenAI, that's changing fast. AI agents are now building a new class of output, what he calls 'artifacts,' and these aren't just documents — they're active, functional tools that unlock a 'far, far bigger' market than most realize.
AI Agents Build Functional 'Artifacts'
Forget the idea that AI mainly generates text or data. Nathan points to a new breed of AI output: artifacts. These are dynamic, living assets, not static reports. “One of the big pushes that we made for this launch was like artifacts, right? Like, both on the model side... and then also on the product side,” Nathan explains. Think beyond a generated spreadsheet to an actual functional website created by an AI agent from a simple prompt. He says, “the canonical artifact that was previously a DAC or something is now becoming a site and like with a site you because it's just HTML you can like it's infinitely flexible.”
This isn't just a formatting tweak; it's a shift in how work gets done. Instead of a team collaborating on a presentation, an agent could build an interactive dashboard that updates in real time. Instead of a developer spending days on a landing page, an agent could prototype a functional one in minutes. These artifacts fundamentally change what collaboration looks like and how individuals manage complex tasks that once required specialized tools or entire departments. It moves from generating information about a problem to generating a solution to a problem.
The "Far Bigger" Market Beyond Early Adopters
Initial enterprise adoption of AI stumbled on a simple question: what exactly do we use it for? Nathan admits, “In enterprise, I think a big part of that is like actually meeting the users where they are like what use cases were they trying to solve and then actually teaching them how they can use AI to like gain leverage there.” Early adopters were quick, but the mainstream needed clear, tangible value. That value is now emerging.
What’s driving this growth? Agentic models are revealing use cases far beyond the obvious. It’s not just about automating repetitive office tasks; it’s about personal automation too. Nathan shares an anecdote: “I've seen people do things in their personal lives that you wouldn't classify as like work technically but like these agents are you know super capable for like one one recent example that someone posted about on our slack is like someone had like a missed package... figured out exactly the apartment complex in which the package was.” This kind of capability, applied across personal and professional domains, suggests the market for AI-powered productivity is exponential. As Nathan states, “I think the adoption is there and and growing fast, but I think the opportunity is like far, far bigger than that. That's where we want to play, especially with ChatGPT Work.”
ChatGPT Work, merging ChatGPT and Codex capabilities, aims to be an AI super app. It’s democratizing the 'magic of code' by letting anyone create sophisticated, functional outputs without needing to write a single line. This opens up entirely new frontiers for productivity, allowing individuals and small teams to manage complex workflows previously out of reach.
What to Do With This
Pick one internal process this week that feels overly complex, involves multiple tools, or requires a specialist. Instead of building a document or a slide deck to explain it, challenge yourself to use an AI agent (like ChatGPT Work) to create a functional 'artifact.' For example, if your team currently compiles a weekly data report in a spreadsheet and then manually pulls highlights for a presentation, try to get an agent to generate an interactive web-based dashboard that visualizes the key metrics and allows for dynamic filtering. Your goal isn't just to automate a step, but to replace a static output with a dynamic, usable tool.