From AI to agency: Puneet Chandok’s 7 rules for the future of work

“From the time Thomas Friedman declared the world to be flat to a post-globalization era, so much is happening. The world seems to be rearranging and rebooting itself,” said Puneet Chandok, President, Microsoft India & South Asia, Microsoft, as he began his keynote speech titled ‘Future of Work and Leading in the Age of AI’ at TechSparks 2025. 

He added that today we’re in the largest infrastructure build in the history of humans, with almost a billion dollars being spent on AI infrastructure every day. “Deep learning is a gift that keeps giving. Models are getting smarter. It is a golden age of software design. There’s a lot going on and the world does seem unfamiliar. This is not something to be worried about. This is something to understand deeply.”

As AI continues to transform the very fabric of work, what will the jobs of tomorrow look like? How will founders and leaders navigate this new era? Which skills will become critical? What will the combination of creativity and machine intelligence unlock? Chandok shared seven predictions on how AI will reshape organizations, jobs, and industries in the future. 

Boundless intelligence

Chandok’s first prediction was around unmetered intelligence, a future where AI is an infinitely scalable, continuously available resource that is as ubiquitous as electricity or the internet. Intelligence today is a scarce and hard-to-access commodity; it requires hiring a team and putting them to work. He said every business on the planet today is a bundle of intelligence and expertise. “However, if these become a commodity and easily accessible, it will rewrite how we run our businesses. Intelligence is not just prediction, thinking, ideas and judgement; it’s also action, and that's coming soon,” he said.

A new digital deskmate

The real unlock on AI will happen when it transforms from a tool on smartphones to a teammate at the office. Chandok called this phenomenon “digital colleagues”, saying, “We’re used to physical colleagues. However, we’re entering a world where none of us will be individual contributors. Every knowledge worker will become an agent boss.”

Three patterns are already emerging to signal this transformation. The first is the presence of copilots, assistants, and the automation of repetitive tasks at the workplace. The second pattern is where agents will work for humans. These agents will have perception, cognition, some intelligence and agency. They will act on the human’s behalf, with their permission, but without their involvement – a generation of autonomous agents humans will see very soon. The third pattern is where agents begin to work, with human supervision, but complete autonomy.

Companies will have a limitless generation of digital colleagues working for them. “If you’re starting up now in the next few quarters and years, I can promise you, your co-founder will probably be a machine and it will work tirelessly, ideate with you, and even send invoices when you’re sleeping,” Chandok said. 

Unbundling and recreating jobs with AI

Prediction three revolved around the billion-dollar question: What will happen to jobs? Chandok shared his belief that AI won’t steal jobs. It will, however, dissect and unbundle jobs. So far, humans have worked hard at managing a bundle of capabilities under a job role. However, this will change. Jobs will soon be like a playlist, where employees can add projects or remove tasks.

As AI takes over different responsibilities, drudgery, and transactional tasks, it will fall to human employees to bundle themselves even more. “If you go back to work tomorrow, what’s the first task of your job that you will unbundle and give away to AI?” he asked. 

From sci-fi to reality: Human in the loop 2.0

Chandok also busted the myth of ‘AI talking to AI’. This is becoming all too real across workspaces today. AI is becoming a sales and support script. While humans are used to the front office being run by humans, while the back office is automated, Chandok shared that this equation is flipping. As an example, he asked the audience to envision how an AI agent could - in the future - call car companies on behalf of an individual, negotiate the model, price, support, package and delivery of the car the human wants, without any involvement. In the event of agents taking on more autonomy (with a human’s permission), humans will turn to edge cases, handling anything that requires real judgement, ethical dilemmas and tough challenges. “This is the human in the loop model 2.0, where we will act as arbiters, not at office, not as operations. We will become quality amplifiers,” Chandok declared.

AI to the rescue

A passionate fan of Batman, Chandok framed the next prediction through his favourite comic book hero. Using Gotham City’s bat signal as a reference, Chandok shared that the startup world is sending its own signal about GPUs. He shared that 15-16 years ago, in the early 2000s, cloud had just emerged and was the focal point of every conversation between startups. Today, founders are asking “what models should i build” or “which API call do i start with?” The next generation of startups will not be built on cloud infrastructure but AI infrastructure. While cloud may scale efficiently, AI brings skills, smarts, and scale to the table. Microsoft believes that the next generation of startups will require three things: agents, copilots and human ambition. 

Transforming the economy

Chandok posited that in the future we will see the death of the “inefficiency economy”, trillion-dollar industries built on inefficiency and greed. For instance, a doctor that requests too many tests or a lawyer that drags out the process in billable hours, all to keep padding their bottom line. AI will bring an end to that charade. Be it proofreading legal documents or decoding conversations with greedy doctors, it will trim the fat. For SaaS companies, this will be how they identify the real Total Addressable Market (TAM). The only metric that will really matter is a closed loop workflow completion. “The next generation of startups that are being built from India will not sell hours, they will sell outcomes,” Chandok said.

Learning: The new competitive edge

Chandok’s seventh and final prediction focused on learning. While previous generations looked at education as a one-time endeavour, this will soon change. The very nature of jobs and tasks is changing every 30 to 60 days. “Learning is now like cardio. You have to do it every day. The industrial age template that we have been working with - learn once, work forever - is gone,” he said.

Acknowledging the job insecurities plaguing employees today, he shared that the real “pink slip” won’t be automation; it will be a refusal to adapt and learn. He advised the audience to bring AI into everything they do, to infuse it into every task. “Build with AI and get it to work for you like a digital colleague.”