Wait, Your Non-Technical Team Can Now Use AI Agents Like an Engineer
If you run a small business in Malaysia, you probably wear a hundred hats. You’re the sales guy, the operations manager, the IT support, and the person who unjams the printer.
Now imagine giving your office assistant the power to instantly automate a task that would take them four hours to do manually. Or having your finance person build a tool to consolidate your monthly data. Sounds like science fiction? A brand new economic research paper from OpenAI just proved it’s already happening inside their own company.
What Happened?
OpenAI released a study on how their AI agent, Codex, is actually being used for productive work. The headline finding is this: agents aren’t just for answering questions anymore. They are becoming the primary tool for every type of worker, from software engineers to the legal team.
Here are the numbers that should grab your attention:
- Massive time saver: Over 80% of individual users now give Codex tasks that would take a human more than 30 minutes to complete, and 25.6% have given it tasks that require more than 8 hours of human work.
- Non-developers are leading the charge: The number of non-technical users has grown 137-fold since August 2025. Your office admin falls into this category, not just your IT guy.
- Everyone is using it: Inside OpenAI, agents now account for 99.8% of the weekly output tokens. The HR team, the finance team, everyone. Their lawyers now generate 85% of their AI work on agents.
- Crossing boundaries: Non-technical employees are using Codex to do coding, automation, and debugging—tasks that used to require a specialist degree.
“Over one-fourth of work done with Codex by workers in business functions was engineering or coding related.” – OpenAI Economic Research Paper
Why This Matters for Your Business
Let’s bring this down to earth. You run an SME. You don’t have a massive IT department. Your team is tight, and everyone already has too much to do.
This paper proves that you don’t need a developer to get the benefits of deep automation anymore. You just need someone who understands the business problem. Here is what this looks like for you:
1. Let Your Finance Team Write Code (Without Learning Code)
Struggling with messy reports? You can ask an AI agent: “Go through our past three months of bank statements in PDF, find the recurring subscription fees, and put them in a spreadsheet in order of cost.” The agent will do the manual grunt work. This kind of complex, multi-step task wasn’t something you could just delegate to a chatbot six months ago.
2. Turn Your Customer Service into an Operations Machine
Agents can handle “long-horizon” tasks. Instead of typing a reply for every refund request, you can ask an agent to: “Check all open refund requests, look up the order status in our system, and draft a resolution email for each one.” It works on it in the background while you handle the customers in front of you.
3. Supercharge Your Marketing Efforts
Your marketing assistant can run parallel agents. One agent drafts social media content based on your brand voice. Another scrapes competitor pricing. A third prepares a report on trending keywords. Together, they do hours of work in a single afternoon.
The main takeaway? If a lawyer at OpenAI can use an agent to build a tool, your operations manager can absolutely do the same.
The Bigger Picture
We are moving from an “on-demand” AI world (ask a question, get an answer) to a “delegation” world (assign a project, get a result).
It feels like the early days of smartphones. At first, they were just for calls and emails. Then they became the primary tool for running entire businesses. Agents are on the exact same trajectory.
The businesses that will likely win in the next few years are the ones that teach their best non-technical people how to act like conductors of an orchestra, rather than just musicians playing one instrument.
This isn’t a future trend. Based on this paper, it’s already the primary way OpenAI operates today.
Ready to figure out how to equip your team?
Book a free 15-min call to see how AI agents apply to your business →