One year later, he says the risk paid off. What began as a bold internal decision has turned into a real-world experiment showing what can happen when a company shifts from human-led operations to an AI-first structure almost overnight. The story has since become a reference point in wider debates about automation, cost-cutting, and the future of work.
Dukaan’s shift from people to machines
The company at the centre of the debate is Dukaan, an Indian start-up that helps small businesses launch online stores. Founder and CEO Suumit Shah had already invested heavily in automation, but in mid-2023 he took a step many leaders still avoid. Roughly 90% of staff were laid off, and much of the customer support workload was handed to AI chatbots. The aim, according to Shah, was direct and uncompromising: slash operating costs and speed up responses. Overnight, human agents were replaced by software, turning Dukaan into a live test of automation at scale.
Immediate backlash and ethical concerns
The announcement sparked instant criticism. Former employees, industry observers, and ethics specialists questioned whether a company built by people could so quickly replace them with algorithms. For many, the move symbolised a growing fear that efficiency gains are being prioritised over human livelihoods, especially in fast-growing tech firms.
Founder claims customer-facing gains
Despite the criticism, Shah has continued to defend the decision. A year later, he describes the outcome as “positive”, particularly in areas customers interact with most. From his perspective, the shift delivered faster service, lower costs, and fewer delays during peak periods.
Performance changes highlighted by Dukaan
- Response times fell from nearly two minutes to almost instant replies.
- Issue resolution dropped from hours to just minutes for common problems.
- Round-the-clock availability became possible without shift planning.
- Operational expenses linked to salaries, training, and offices declined sharply.
- Automated queues replaced long waiting times, with humans stepping in only for rare cases.
Shah argues customers value speed above all else. In his view, lower costs combined with faster support create a stronger and more sustainable business.
Critics say speed misses the bigger picture
Opponents counter that metrics like response time do not capture everything. They raise concerns about accuracy, empathy, and long-term trust, questioning whether these qualities can be measured as easily as minutes and seconds.
The wider fear of AI-driven job losses
The Dukaan case feeds directly into a long-running debate: what happens to work when AI becomes cheap, powerful, and always available? For many observers, this single company reflects a much larger shift underway across industries.
Two competing views on AI at work
Reactions generally fall into two camps. Supporters see AI as a support tool that handles repetitive tasks like password resets, shipping updates, data entry, and standard FAQs. In this view, automation frees humans to focus on complex decisions, creative work, and relationship building.
Critics, however, see a creeping employment crisis. They argue that once AI is “good enough,” companies may remove workers entirely rather than assist them. For them, Dukaan is less about innovation and more a warning about how quickly white-collar roles can disappear.
What this experiment really shows
Dukaan is not the only firm using chatbots, but few have moved as quickly or as decisively. That makes it a useful case study, even if it represents just one data point in a crowded field.
Why customer support was an early target
Customer support has long been suited to automation. Many requests are repetitive, rule-based, and tied to structured data. Tasks like order tracking or refund checks fit AI well, while situations involving emotional distress, negotiation, or ethical judgement still favour humans. Seen through this lens, Dukaan’s faster resolutions are unsurprising. The harder questions concern displaced workers and corporate responsibility.
Efficiency versus livelihoods
AI decisions do not exist in isolation. Choices like Shah’s sit at the crossroads of profit pressure, social expectations, and real lives. Supporters argue that refusing automation in a cut-throat e-commerce market may be financially reckless. Critics respond that rapid layoffs can devastate communities, especially when retraining or transition support is absent.
What an AI-first workplace might look like
Beyond Dukaan, many industries are quietly testing AI-first models. Instead of adding tools to human teams, they start with software and introduce people only where needed. In such setups, AI handles most communication, reporting, and routine decisions, while humans focus on strategy, vendor relations, and high-risk disputes. Headcount shrinks, roles become more specialised, and pay often rises for those who remain.
Hidden risks behind simple labels
Terms like “chatbot” and “job displacement” often mask complexity. Modern chatbots connect directly to databases and payment systems, allowing them to act, not just reply. That power means errors or bias can scale rapidly. Job displacement, meanwhile, can be gradual or sudden. In cases like Dukaan, speed is the shock factor, leaving little time for workers to adapt.
If you want to avoid loneliness at 70 and beyond, it’s time to say goodbye to these 9 habits
A choice more companies will face
The Dukaan experiment will not be the last. As generative AI improves and costs fall, more leaders will confront the same dilemma: pursue maximum efficiency or retain more people than financial models demand. Where each company draws that line will shape not only its balance sheet, but the future structure of work itself.
With its 337 metres and 100,000 tonnes, the largest aircraft carrier in the world rules the oceans
