AI Leads To Job Cuts In Customer Support: But Will Push For Agentless Setup Hold? Experts believe that AI deployment can help companies save nearly 50% or even more on costs.

AI is fast replacing humans in the customer service departments of several large corporations. And, several new-age companies are now having AI-driven customer support departments with no or bare minimum human interventions.

In the last couple of years, we have seen several job cuts or the so-called restructuring affecting the customer service division. For instance, Salesforce reduced its customer support team by 4,000 employees or even higher. Its executives said that AI-driven systems were now handling nearly half of the customer interactions. Amazon, Accenture and TCS too have let go of people in favour of deploying AI-based tools.

Limechat cofounder Nikhil Gupta recently said that their AI agents can help their clients reduce the work to handle customer queries by nearly 80%. Zomato and Paytm are some of other top names to have restructured their workforce to align with AI.

Economic Advantage and Deployment

Cost. That's a common argument for reducing the workforce in the service department. Experts believe that AI deployment can help save nearly 50% or even more on costs. There are also big savings in management time and context switching, since AI can handle repetitive queries efficiently and free up humans for complex tasks.

RevRag.AI cofounder and CEO Ashutosh Prakash Singh breaks it down: "A human support agent typically costs around INR 8–INR 12 per minute, while an AI agent can cost between INR 1–INR 4 per minute, depending on the level of naturalness and customization. This translates to a cost reduction of 50% to 90%, depending on the scale and complexity of deployment. Beyond savings, AI also increases efficiency by handling multiple conversations simultaneously, further enhancing ROI."

Experts add that the primary economic advantage comes not only from reduced headcount but also from response-time optimisation, 24/7 availability, and knowledge retention. AI systems with memory-based architecture which further reduce training and onboarding costs by continuously learning from historical data, ensuring that every new interaction is contextually aware.

From a tech point of view, deployment of AI is fairly easy or complicated depending upon the company's workflows and anything related.

"Integrating AI into existing workflows, databases, and CRMs is becoming increasingly straightforward thanks to modern integration platforms. However, giving proper context to AI systems—so that they understand business logic, tone, and customer data—still requires thoughtful setup," Singh added.

But will the transition to AI hold?

Humans and AI's limitations

According to a Gartner report published last month, none of the Fortune 500 companies will have fully removed human customer service by 2028.

Additionally, it was predicted that by 2027, half of the companies anticipating a significant reduction in their service workforce due to AI would abandon these plans. This would be mainly because several of them would not achieve their objectives of agentless staffing.

"We expect fewer human agents, but not completely agentless organizations. The most successful organizations are those that balance technology with the human touch, redeploying their teams to focus on growth and customer satisfaction," Kathy Ross, Senior Director Analyst in the Gartner Customer Service & Support practice, said in the report.

It's a fact that AIs are increasingly getting better. They understand the context much better than ever before, and it's plausible that they are going to be significantly better in days to come. And it's safe to guess that the AI (such as agentic AIs) will take over the mundane tasks and so on. However, humans will not completely go away from the larger scheme of things.

Neural Arc founder and CEO Aniket Tapre notes that a visible transition is underway, particularly among technology, fintech, and e-commerce companies.

"However, evolution will not eliminate human support entirely. Instead, it will create tiered intelligence ecosystems where AI autonomously handles up to 90% of user interactions, and humans intervene only when business-critical or emotionally complex situations arise. The long-term trajectory is AI-human symbiosis, not replacement. The companies leading this transition will be those that treat AI as an intelligent collaborator rather than a mere automation layer," he told Entrepreneur India.

Naveen Saraswat, cofounder of Letest AI, adds that it's too early for companies to have bare minimum human interactions as the AI/LLMs hallucinate a lot. Saraswat goes on to say that he'd give another decade to the tech to further improve.

Saraswat's apprehensions are not off the mark. Modern AIs do have a problem of hallucinations. Essentially, a lot of models do generate false or fabricated information despite having advanced learning capabilities. They can even conjure up imaginary citations, incorrect facts, and statements.

The most famous and recent example of an AI-led debacle is Deloitte's one of the reports for the Australian government. Both the parties were widely criticised after it was revealed that the 237-page report was filled with references to sources and industry pundits that never existed. Deloitte later acknowledged that parts of the report were AI written.

Evidently, a total reliance on AI is premature.

"I've seen a few companies try going fully AI with customer support — but most eventually rolled it back. From what I've observed, pure AI rarely works alone; there always needs to be a bit of human touch to handle tricky or emotional situations," Ayush Shukla, Co-founder and CEO, Nector.io told Entrepreneur India.

Emotions

As mentioned above, there has to be a right balance between the application of AI and human interventions. Also, AI has become smarter than ever before but its limitations, especially understanding human context, emotions, and nuances is very important. And this is important for customer service.

Tapre explains that AI-driven customer support systems are redefining efficiency and response precision. However, "better" is contextual. In high-frequency, low-complexity queries such as order tracking, password resets, or policy lookups AI consistently outperforms human agents in speed, accuracy, and availability.

"Yet, when emotional nuance or complex judgment is required, human empathy remains irreplaceable. The most successful enterprises use a hybrid model: AI handles repetitive workflows while human experts manage escalations," Tapre said.

On the possibility that AI-driven customer service can be better, Saraswat added that it depended on the fact that the company that has deployed AI support Agents have trained their LLM on years' worth of past support data.

It has been noted that 90% of the customers getting in touch have similar issues which has been reported earlier. This has been the case with B2C companies. However for complex B2B SaaS products the convenience for agents reduces drastically as the issue can only be addressed by a human, he said.

Singh pointed out that AI still struggles with complex, contextual, or emotionally sensitive issues that require human judgment or negotiation. For example, when it comes to a discussion about loan interest rates or personalized dispute resolution, human agents tend to perform better.

"The ideal model is a hybrid system, where AI handles most routine queries while humans step in for complex or nuanced cases," he said.