As reported by LONDON, Feb 20 (Reuters Breakingviews) – Let’s rewind to 1999. Boo.com, an online fashion startup, had a grand vision: a virtual shopping assistant named Miss Boo. This AI-powered avatar promised to give fashion advice and showcase clothes in 3D. Sounds futuristic, right? Fast forward a few months, and Boo.com collapsed. The technology was slow, clunky, and ultimately, a business disaster.
Now, in 2025, tech giants are betting on a similar concept—only this time, they believe AI agents will actually work.
The Rise of ‘Agentic AI’
If you’ve been listening to Silicon Valley’s latest buzz, you’ve likely heard the term “Agentic AI.” The likes of Microsoft, Alphabet, and Amazon have collectively thrown tens of billions of dollars at large language models, hoping these AI-driven assistants will revolutionize business processes. But there’s a problem—Chinese rival DeepSeek has launched a competing AI model that matches Western counterparts at a fraction of the cost. So, what’s the next move for Big Tech? Enter AI agents.
What Makes Agentic AI Different?
Here’s the promise: AI agents won’t just follow commands; they’ll act independently, proactively solving problems with minimal human intervention. Unlike traditional AI automation, which relies on rigid rule-based programming, these new agents will:
- Think and act independently – They won’t wait for instructions but will proactively handle tasks.
- Work toward goals, not just follow steps – Instead of just responding to inputs, they’ll optimize results, like maximizing sales or improving efficiency.
- Handle multiple tasks simultaneously – From retrieving data to making financial transactions, all in one workflow.
Sounds great in theory. But let’s be real—how close are we to this sci-fi reality?
The Real-World Applications
For now, most AI agents are glorified customer service bots. Take Lufthansa’s AI chatbot—built by Cognigy—to handle rebooking when flights get canceled. Airlines spend 5-10% of revenue on customer service, so if AI can absorb that workload without increasing costs, profit margins will improve.
Amazon already claims its AI agents have saved $260 million by automating over 30,000 production applications. Salesforce CEO Marc Benioff went even further, suggesting that AI-driven productivity gains mean his company won’t hire new engineers in 2025.
The Big Bet on AI Agents
Investors are paying attention. Ark Investments estimates AI agents could facilitate 25% of global e-commerce sales—worth $9 trillion—by 2030. Companies are racing to integrate AI assistants into their workflows, whether by building them in-house (like DoorDash and Pets at Home) or buying them from SAP and Salesforce. Uber, Instacart, and Lyft are all inking deals with OpenAI and Anthropic to supercharge customer interactions with AI-powered operators.
But Here’s the Catch…
Rolling out AI agents isn’t as simple as flipping a switch. Companies need to train employees to work alongside AI, fine-tune models to avoid costly errors, and brace for regulatory scrutiny. Financial firms, for example, could face massive losses if an AI agent mistakenly approves high-risk loans. Air Canada learned this the hard way when it had to refund a customer misled by its chatbot into purchasing an overpriced ticket.
Beyond that, AI agents might shake up traditional business models. Legal and accounting firms, for example, rely on billable hours—but if AI speeds up contract reviews and audits, those firms will need a new way to make money. And if junior-level tasks get automated, how will businesses train the next generation of workers?
The Bottom Line
Microsoft CEO Satya Nadella envisions a future where building AI agents is as easy as making an Excel spreadsheet. But here’s the question investors should be asking: will companies spend big on AI agents, or will the return on investment prove elusive?
For now, Agentic AI is brimming with promise—but whether it delivers a tech revolution or ends up as another overhyped trend remains to be seen.
