The arrival of AI agents is turning workplace conversations from speculative to urgent. The focus keyphrase here is AI agents and jobs, and the truth is they are no longer a futuristic thought experiment. From customer service chatbots to semi-autonomous project managers, these digital colleagues are starting to shape how entire industries operate. The question, then, is not simply whether AI will change work, but whether it will change the very meaning of having a job.
AI agents are rapidly moving from copilots to semi-autonomous colleagues, able to manage multi-step tasks,Some firms are reducing headcount, while others are repositioning humans as orchestrators of AI fleets,The future of work may hinge on collaboration between humans and AI rather than outright replacement
From copilots to digital colleagues
Generative AI’s early appeal was in automating content: text, images and video. But the new generation of AI agents goes much further. They can research companies, write blog posts, manage clinical trials, optimise websites and even orchestrate workflows across departments. Salesforce chief executive Marc Benioff cut 4,000 customer service roles earlier this year, stating bluntly that AI agents had cut support costs and reduced the need for staff.
That trend is mirrored in investment. IDC estimates global AI spend will reach $632 billion by 2028, while the IMF believes AI could impact 40% of jobs worldwide.
The rise of the agent economy
Recent weeks have seen a flood of product announcements. HubSpot unveiled 15 new agents and a marketplace. Adobe launched six plus an “agent composer”. Airtop offers a conversational interface to build your own. Korean telco SK Telecom has partnered to build sector-specific agents. Even niche areas such as clinical trials now have dedicated AI agents, courtesy of Grove AI.
Industry analyst Jeremiah Owyang goes further, suggesting we will soon see autonomous organisations run entirely by AI agents. He imagines entities that generate revenue, sell services and operate with no human oversight at all. It is a provocative vision, somewhere between Silicon Valley idealism and a cautionary tale for labour markets.
Echoes of past revolutions
Some argue the AI shift is simply another phase in the evolution of work. Airtop founder Amir Ashkenazi points to the industrial revolution, when farming fell from employing 80% of humans to less than 2%. Jobs did not disappear entirely; they transformed. His company’s conversational builder is designed to “democratise automation”, giving individuals a chance to reclaim some agency in how tasks are delegated.
Ashkenazi himself uses AI to capture meeting transcripts, generate action items and draft follow-up emails. The human remains in the loop: editing, adjusting, approving. The point is not that AI erases labour, but that labour morphs.
“AI is turning everyone into a manager of agents,” he suggests. “Careers will be defined not by how many tasks you can do, but by how effectively you orchestrate a fleet of AI agents.”
“AI is turning everyone into a manager of agents,” he suggests. “Careers will be defined not by how many tasks you can do, but by how effectively you orchestrate a fleet of AI agents.”
Hybrid models of work
HubSpot has taken a similar view, championing “hybrid teams”. Karen Ng, the company’s head of product, insists human creativity and judgement remain irreplaceable. Her colleague Dharmesh Shah, HubSpot’s CTO, describes agents as “force multipliers” designed to enhance rather than erase human ability. Research supports that thesis: one study suggests individuals with AI outperform groups of humans working without it, but teams equipped with AI deliver the best solutions of all.
The new workplace dynamic could therefore become a mix of human ingenuity with machine efficiency. Or, put more plainly, AI will not take your job, but someone using AI might.
Risk and reward
There is, of course, risk. AI agents are probabilistic, not deterministic. They will get things wrong. HubSpot freely admits its tools sometimes fail spectacularly. Humans must remain in the loop to ensure accuracy and quality.
But there are upsides, too. Tasks that are tedious or repetitive can be automated, allowing people to focus on more strategic or creative pursuits. AI also lowers the barriers to entrepreneurship. Ashkenazi foresees “unicorns with fewer than ten employees”, thanks to agents that can handle marketing, operations and product testing with minimal human intervention.
The human mandate
Adaptation is the constant. For all the utopian and dystopian scenarios, the reality is that AI agents are here and becoming more capable by the day. Workers who learn to configure, customise and collaborate with them will thrive. Those who resist may struggle.
The new mandate is straightforward: find the agents that suit your needs, and learn how to use them well. The future of work may not be man versus machine, but humans learning to direct and leverage fleets of digital colleagues.
Brave new world? Perhaps; as history shows, revolutions in work have never ended labour. They have only redefined it.
Latest Comments (2)
It's all well and good for Salesforce to cut 4,000 jobs, but what about the accountability with those AI agents? The EU AI Act is very clear on risk assessment and human oversight. I don't see how this new "agent economy" can possibly function without a clear framework for liability, especially when these systems are making critical decisions.
@AIinASIA Salesforce cutting 4,000 customer service roles because of AI agents... that's a big number. I wonder how much of that is applicable here in Malaysia. Our customer service landscape is so different, often relying on really personalized, multi-lingual human interaction that current AI agents probably can't fully replicate without some serious local training data and cultural understanding.
Leave a Comment