Pratik on Emergent Behavior: growth strategy for an AI in farming

Pratik was on EB-10, the tenth episode of Emergent Behavior, the AI takeoff podcast run out of the Cogniscendo newsletter. The episode dropped on 23 April and the host’s written breakdown went up this week. It is a candid conversation about the operating reality of running an AI product for Indian farmers, the parts that do not show up in launch posts.

The exchange people keep sending us is about what happens when a farmer meets voice AI for the first time. “Farmers would just say hello. They don’t know what to ask,” Pratik says. Some go quieter than that: “Some just started recording but then don’t say anything at all because they are just overwhelmed.” Every product plan assumes the user can phrase what they need. For a first-time voice user standing in a field, that assumption breaks, and you end up designing for the hello.

The growth section is the one I would make founders sit through. “We cannot keep expanding the user base. I’m trying to kind of mellow down the growth on the consumer side before we actually start generating revenue.” Almost nobody says that publicly. For a smallholder product, more usage at zero revenue is a worse problem, not a better one, and pretending otherwise just burns the runway faster.

The reason the revenue cannot come from the user is in the episode too: “Farmers cannot be monetised in any developing nations. It’s mostly the businesses that you have to work with.” That is exactly how we sell today. The farmer talks to the product; the agribusiness pays for it.

Two shorter threads are worth flagging. Agriculture in India is a state subject, and Pratik is blunt about how much politics moves through it with every change of state government, a piece of context most “AI for India” pitch decks skip. He also describes the distance between the people doing cutting-edge research on climate-resilient seed and fertiliser and the farmers who are supposed to benefit from it, which is the gap domain models exist to close.

Near the end he sketches a farmer with a personal drone, taking a picture of the field from the sky every few days and asking the AI what is going on. Dhenu Vision, which we shipped in March, already covers the ground floor of that idea: put an image in front of the model, ask, and get an answer keyed to the crop. The sky part can wait.

The full episode is worth the listen. The host’s writeup, with the rest of the quotes and the production notes, is at emergentbehavior.co. The Cogniscendo newsletter is on Substack if you want the weekly cadence of AI takeoff coverage.