AI as your tireless co-founder: transforming operations, augmenting human talent, but still needing a human touch to drive success.
getty
What if your next co-founder never sleeps, never asks for equity, and can code, analyze markets, and manage customers all at the same time?
For startups operating on tight budgets, agent AI promises just that—intelligent systems capable of conducting market research, writing code, managing customer communications, and optimizing supply chains in real time. Two years ago, this sounded like science fiction. Today, it’s a living experiment that forces entrepreneurs to confront what it really takes to build a business.
The race to replace workers is already on — but is anyone really winning?
Swedish fintech Klarna has emerged as one of the most ambitious testers of this technology, reducing its workforce from around 5,000 employees to 3,000 through its strategic deployment of artificial intelligence. The company’s AI assistant now handles the workload of 700 customer service agents, handling 2.3 million conversations in 35 languages in the first month alone.
said CEO Sebastian Siemiatkowski Bloomberg“I’m of the opinion that AI can already do all the jobs we humans do,” while acknowledging the nuances of the statement. In mid-2025, Klarna started recruiting human customer service agents again. Siemiatkowski explained that while AI customer service was more cost-effective, it produced “lower quality” results and that it’s “critical to be clear to your customer that there will always be a human if you want them.”
Klarna’s example shows what mature AI development looks like: aggressive testing, rapid iteration, and intellectual honesty to adjust course based on results.
While Klarna demonstrates the potential of AI in customer service, other AI systems are now tackling a much wider range of business tasks, from coding to data analysis.
What AI agents can actually do right now
The OpenAI operator is an artificial intelligence agent that can interact with websites on its own—filling out forms, making purchases, and completing tasks as a human would. It understands what’s on a screen without the need for special APIs, making it capable of handling repetitive web tasks like ordering groceries or automatically generating content.
Microsoft has integrated AI agents into the Microsoft 365 Copilot suite, assisting users in Word, Excel and PowerPoint. These agents can automate routine work, create content, summarize meetings and highlight key action items – freeing people to focus on higher-level thinking.
Maybe another smart one is Cognition LabsDevin is an AI software engineer who can design tasks, write code, test and debug software, and even develop it — all autonomously. Devin handles routine coding tasks so human developers can focus on designing systems, solving complex problems, and creative innovation.
The real risk isn’t the takeover of AI — it’s premature development
McDonald’s stopped the AI drive-thru system after repeatedly adding McNuggets to orders until it reached 260. Air Canada was ordered to pay compensation after the chatbot gave him incorrect bereavement fare information. These weren’t malicious AI failures—they were systems that were growing faster than organizations could properly test and manage.
of MIT “The GenAI Divide: State of AI in Business 2025” shows that approximately 95% of AI production pilots fail to deliver measurable returns. Failures are often due to poor integration with workflows, wrong focus, over-reliance on internal development, lack of adaptability and gaps in governance and oversight.
Together, these examples highlight that mature development is necessary for the successful implementation of artificial intelligence.
The Verdict: Can Agentic AI Build and Run Your Company?
AI agents can perform critical functions and streamline operations at unprecedented speed. However, true and lasting success still depends on human judgment, creativity and strategic oversight. The real winners will not be companies that chase automation for its own sake, but those that master the art of human-AI collaboration—organizations that design intelligent ecosystems based on vision, adaptability, and mature growth. These are the businesses that will transform AI from a tool to a competitive advantage, shaping industries rather than simply reacting to them.
The question is no longer whether AI can replace humans, but who will lead this new era of intelligent business.
