A speaker (from a consumer-company context) explains how AI and data (an internal GPT inside a data lake) are being used to improve supply-chain decisions and demand forecasting for perishable consumer products. The speaker contrasts past, history-based demand programming with current AI-driven, store- and region-specific predictions (examples: Merida vs. Chiapas) and notes earlier demand "assertiveness" was about 60%, causing large unsold-perishable waste; AI now provides granular signals (what customers buy at specific stores/events). The talk also mentions emerging concerns about autonomous agents organized with supervisors.
Questions this video answers
Approximately 60% assertiveness in demand.
Chapters
The Internet. The Internet of everything. Data, artificial intelligence. Here you have some little monkeys called agents. That's scary. There goes Bafar. I mean, we're developing agents, but those agents then have an organization chart and a boss. And those agents start to learn and start to make decisions. That's really scary. Where is this going? No. Oh, dude. Let's see. The supervisor is Agent 002, man. So. Oh, great. And then that's what's happening. That's where we're going. And well, where we are right now. What can I tell you? For example, the supply chain, which has always been a huge issue, right? Being able to define demand. Imagine our products are perishable products. Imagine programming 40 plants with the demand that you think you're going to sell. So this has changed in a very interesting way. Maybe before we used to do demand programming based on how you see it and what the history is and all that stuff. And you had your data engineers. Yes, but the history is no longer useful because if food made an offer, then you're left with Virginia Buffaro. If San Rafael, then Sabores, you can't control that. So to define demand, we had 60% assertiveness in demand. Imagine the refrigerators in Merida and Pueblo with 40% of the product that you didn't sell. Perishable. Well, it was chaos. Now the systems can tell you what is being bought at Suriana, the rodeo or the bullfighting or whatever you want. What is the lady buying? And what does the system tell you about who is eating it? In other words, this house is buying flavored ham and turkey breast and it is sold at Superama, I don't know which one. And on this side they are buying this. And in Merida it is different because they eat differently. And in Chiapas it is a different process. All that is being given to you by artificial intelligence. All that is telling you what direction you are taking in the lake. Let's go where we are. We have an internal GPT chat, right? So all this is inside the lake.
