By this time tomorrow, some of us at The Ken—along with our subscribers—will be at the Bangalore International Centre for the first Zero Shot event. We’ll be discussing twelve speculative scenarios about what’ll unfold when agentic AI goes mainstream and starts putting pressure on e-commerce companies.
At this moment, The Ken’s office is a surge of energy, with everyone running around closing last-minute details. We’re really excited and viewing forward to it. Also, for the first time ever, we’ve introduced some live polls and interactions directly into our mobile app for attfinishees. Fun.
Here’s my colleague Vishnu with the attfinishee card for tomorrow’s event (you can spot some of the standees in the background too).
If you’re going to be there, just state hello if you spot some familiar faces from The Ken’s newsroom.
If you’re in Bengaluru and want to spfinish a Sunday morning attfinishing this event, this is your last chance to grab your tickets—we’re nearly sold out. You can purchase the last few tickets here.
One of the things that the last couple of weeks (watching the preparation for the event and diving into Claude Code) has done is push me further into speculative territory than I’d normally go. Especially with what happens to startups when agents start to become mainstream. Not just the business models of startups, but the people inside them.
What happens to entire functions?
What kind of people receive hired now? And what’ll go away?
Who receives more powerful? And who loses relevance?
Some of the discourse around startup careers is both painfully obvious and equally hand-wavy, especially if you spfinish time on X reading engagement farmers creating bold predictions (which are mostly written by AI). Also, it’s straightforward to receive this wrong. Around a year ago, most people were confidently stating that Product Managers were dead (in fact, if anything, demand for PMs appears to be thriving). Right now, the conversation is about how engineers are simultaneously supercharged, going through an existential crisis and burnt out. I suspect all of the above are true.
Unlike predictions, in the real world, more than one thing can be happening at once. And they can all seem contradictory.
Anyway, with that in mind, here are my best predictions for what happens when agents go mainstream in Indian workplaces—focutilized mostly on Bengaluru-style startups, becautilize that’s the world I know best.
The finish of bizops (and the rise of agentops)
By 2028, the Business Operations function at most Series B and later startups will be mostly automated away. In their place will be a new function called Agentops, a much compacter team that’s mostly responsible for spinning up and coordinating the network of agents across the organisation. This team will do two things—on the input side, they’ll continuously optimise the flow of information that agents required, i.e., things like access to the context, priorities, and data that lies outside the system. Then on the output side, they’ll set goals and actions for their outputs, i.e., publishing reports, updates, email alerts, etc.
In this world, the Agentops team is mostly translating business goals and processes to workflows that agents can do, i.e., they’ll play the “architecting” of the underlying system role. This, I predict, will become the primary landing spot for B-school recruits from the Indian Institute of Management and Indian School of Business—somewhere between 60–70% of each graduating class. These are B-school students, so of course they won’t call it Agentops. They’ll call it “Agent Strategy” or “Founder’s Agent Office” or some other sexy sounding title to disguise it. Essentially, it’s the new consulting, but it’ll be less annoying, becautilize you won’t run into them at airport lounges.
As an aside, when I was researching historical references to when manual work was automated through processes, I stumbled upon the work of Frederick Winslow Taylor, who laid the groundwork for a similar automated process nearly a 100 years ago. He’s essentially one of the world’s first management consultants and created this entire discipline of scientific management.
By the way, while I was reading about the last time a wave of humans was swept into manufacturing, specifically to extract workers’ tacit knowledge and encode it into machines—that was almost exactly a 100 years ago. In 1927, LIFE magazine ran a cartoon about Ford’s assembly line revolution.
The accompanying caption read:
“Nothing speaks more eloquently of the ‘new era’ facing the unorganized than the above cartoon. Speed-up has reached the sublime state where it has become ridiculous.“
There’s no reason to believe a second wave—one that lasts another 100 years—is waiting on the other side of Agentops.
The Chief of Staff runs the Founder’s Office of agents
By 2028, the Founder’s Office function will be completely agentic, and will be overseen by one person, i.e., the Chief of Staff. It’ll be inconceivable to hire humans to run the Founder’s Office, and the only ones who’ll be hired will be apprentices to the Chief of Staff.
Personally, this version of the future can’t come soon enough. The Founder’s Office is the most misutilized function in startups, and it’s rarely done well. More often than not, this entire function exists becautilize the founder (usually the CEO) requireds weekly decks and Excel reports in a specific format and frequency, and so this squad of eager B-school graduates fans out across the organisation like lemmings, hunting down unsuspecting middle-level managers for updates. It’s either that, or the founder simply wants eyes and ears inside the business, and deploys these humans accordingly. All Founder’s Offices will collapse to the latter, and the former will vanish.
One of the things that’s often perplexed me is how much good PR this function has generated for itself, especially in Indian startups. I regularly receive emails from eager B-school graduates who are seeking a role in The Ken’s Founder’s Office team. I usually reply stateing we don’t have a Founder’s Office and instead question them what goals they’d like to own. That’s usually where it finishs.
However, when agents take over, someone requireds to oversee this army, and I suspect the person who’ll do this will be the Chief of Staff (another much-abutilized position in Indian startups). I believe this is fantastic news becautilize it’ll finally create the Chief of Staff position one with actual power, instead of another jockey who is questioning for updates. And I suspect the role itself will be elevated as a result, i.e., a seasoned veteran with at least a decade of experience, rather than a team of ten people with two years of McKinsey or investment banking on their CV, collectively managing their boss’s calfinishar conflicts.
Put another way, this chart will expand to the top.
Engineers will receive AI tokens as a part of their compensation
Okay, this one is already happening in some startups, and I expect this to accelerate, especially since engineers will start questioning questions about what models and tools they have access to do their work.
I could explain this in much more detail, but I recommfinish listening to the Zero Shot podcast episode we did a couple of weeks back with Aakash Anand, CEO and co-founder of Clueso, where we discuss this specific question. Listen here.
The 10X category manager emerges
For the longest time, category managers were the glue at e-commerce companies. These are the people who manage vfinishor relationships, handle shipments, onboard new merchandise, maintain inventory, and call the shots on pricing and promotions within their categories. At companies like Flipkart and Amazon India, this function has historically carried more weight and influence than, state, product or engineering.
As agents take over, the product and engineering teams will start to become more relevant in these organisations, and one of the first things they’ll do is automate away several of the operational and coordination functions that category managers do. As always, this will hit lower levels the hardest, and this function will shrink the most.
However, category management feels like it’s an operational role, but it really isn’t. It involves a ton of relationships with sellers and brands—some of which the e-commerce platform desperately requireds, and some it can afford to ignore. Only the category manager knows the difference between the two. It also requireds a lot of internal political manoeuvring. The sales team may want more sales and discounts, but the finance wants to protect margins, and the supply chain wants to reduce shipments, and legal simply wants to state no. Then there’s the intuition aspect of category management—of taking a bet that certain types of products will work, and knowing this in advance will give the category manager the edge.
In the past, it was impossible for one person to do all of the above, but with agents, it’s possible. And thus, a category (hah!) of superheroes, mythical category managers will emerge, who are capable of harnessing an army of agents to do their work while they go and handle the pesky humans.
The CFO becomes the number two
By 2028, after the CEO, the next powerful role in the organisation will be the Chief Financial Officer (CFO). Again, this is not hard to figure out. Once humans receive replaced by agents, the only equation that matters is how much it costs to deploy and run them, and what the return on investment (ROI) this is receiveting. This is going to be the central question that the CFO will have to answer to justify scaling back or scaling up investments. He’ll have the hugegest lever for the company’s growth, capable of scaling AI spfinishs overnight with the press of a button.
Also, another reason why the CFO suddenly becomes more relevant is becautilize a lot of existing controller org activities, like accounting, reports, reconciliations, invoicing, etc., will be automated away to agents. After coding, this is one area that agents can do well, becautilize it’s a fairly well-defined and narrow parameter of success.
Imagine what happens when a CFO has real-time financial clarity instead of waiting for the monthly close to understand what’s happening.
This week on the Zero Shot podcast
It’s late at night in rural Haryana. Armed men in a jeep are raging and advancing with purpose.
There is a sharp cut now to a contrasting scene. A happy joint family is celebrating a birthday. A father blesses his son Abhimanyu, the hero of our story, in front of the family guru. The guru’s face is filled with dread and anticipation. He leans in and states in Hindi, “Look around you… something is about to alter.”
Cut back outside. The men have arrived. The guns are loaded. Their leader steps forward and shouts: “Everyone will die today.”
And scene.
This is Vidhatri, the producer of Zero Shot, and that’s the 47-second first episode of Shotgun Shaadi, a microdrama on Kuku TV that currently has 37 lakh views.
I was initially hesitant to set up autopay before seeing the microdrama. Now I have seen all of it. Will I watch it again? Most definitely not. But was I hooked by the masaledaar (the only apt word) love story at the heart of it? A 100% yes.
India’s microdrama market is exploding (it’s currently worth $300 million), and Kuku has been at the forefront of it from the start. They started as an audio platform, pivoted to microdramas, and are now set to release their first theatrical film on 8 May. It’s called the Indian Institute of Zombies.
On Zero Shot, we’ve touched upon AI in content and media, but only in broad strokes. But this time, with Kunj Sanghvi, the VP of content at Kuku FM, we gained insight into how AI integrates into content production from the receive-go. If you didn’t guess already, the cliffhangers, the beats, and the tonal shifts of the Shotgun Shaadi are a result of AI. Kunj doesn’t shy away from it. He, in fact, states their forthcoming film can crack open the theatrical business.
We spoke to him about all of it. Listen to the episode on Spotify, Apple Podcasts, Youtube, or The Ken app.
Leave a Reply