Ganesh Subramanian

Work+fun @stylumia

How can multiply outcomes using decision agents ?

Move from data - dashboard to decisions !

Predicting potential demand of a new product at idea stage is now a possibility!

One of the most valuable skills today is believing that things get significantly better over time. This is especially true in technology. When I was at Myntra in the early years, buying clothes online in India felt like a leap of faith. Sizes were uncertain, delivery took days, and returns were a hassle. Many smart people looked at that moment and concluded it would never work at scale. A few years later, it became how millions of people shop. I see the same pattern with AI today. The early versions make mistakes. They hallucinate, they miss context, they need hand holding. It is easy to look at that and form a fixed opinion. What has helped me is treating my views on any early technology as temporary. What I observe today is a snapshot, not the destination. The real question is not how good it is now, but how fast it is improving.

By the time you see a trend in a traditional way, you have missed a few months ...

It is the domain expert who builds. I have spent many years in retail. And I see brilliant domain experts everywhere playing one role: advisor. Sharing wisdom in meetings, reviews, calls. Advisory is valuable. But it has two problems. It scales 1:1. Your knowledge helps one team, one project, one conversation at a time. It does not self-heal. The world is changing so fast that advice given today decays quickly. A built system learns, improves and compounds. For the first time in history, domain experts do not need to code to build. The tools have crossed the threshold. What took an engineering team now takes curiosity and a weekend. Domain expertise plus building capability is the exponential combination of this decade. Experts who build will create impact that advisors simply cannot. I have been testing this belief with a free community course, 100 Days of Agent Building. Four cohorts done. Hundreds of non-coders in retail now shipping working AI agents. Cohort 4 ends in 2 weeks. Cohort 5 is opening in October. Our alumni are all over the world from US to Australia. If you are in retail, any function, non-coder, and want to move from advising to building, apply here and join the global community building #futureskill lnkd.in/eidyakaz The best time to be a domain expert is now. But only if you build. #AIAgents #NoCode #RetailTech #DomainExperts #100DaysOfAgentBuilding #FutureOfWork

Shelves Lie, Baskets don’t ! See you at @NRFnews Paris !

Retail intelligence is reimagined as as action engine - data to decisions …

Ten years ago Stylumia began with one belief: retail decisions deserve a better signal. Today that belief becomes Orbix. Officially live. 4 AI agents. 1 digital twin of your market from 350,000+ brands. 4 decisions: what to launch, stock, back, price. Retail decisions, reimagined. orbix.stylumia.ai

We are hiring for one of the rarest minds in AI right now. Not a specialist who guards one box. Someone who can hold the whole thing in their head. Raw data to living product. And keep the quality bar high at every step in between. The best agentic AI companies learned this the hard way. A demo is easy. A system that stays reliable in production is the real test. And the people who pull that off are almost never the ones with the longest resumes. They are the ones who think sharp and learn fast. Agentic AI is barely one to two years old. The field is still being written. So I care far less about years of experience and far more about how you think. The person I am looking for: You see the whole system, not just your slice. You have built something agentic from raw data to a working product, and felt where it breaks. You treat data quality as a first class problem, not an afterthought. You learn faster than the field moves. You can say I do not know yet, then keep digging until you do. If this is you, or someone you respect, reach out. The work is real. The problems are hard. The bar is set where the best in the world set it.

What Elon Musk’s Antimatter Tweet Reveals About Retail’s Future Recently, @elonmusk posted a tweet that I found surprisingly thought-provoking. He suggested that one day humanity might spend “a trillion times a trillion dollars” producing antimatter to travel to other star systems. When someone questioned the scale of such a number, he replied: “Things won’t be measured in dollars then, just mass & energy.” I’m certainly not qualified to comment on the future economics of interstellar travel. But the tweet stayed with me. Not because of antimatter. Because of what it implied about leverage. A few grams of antimatter can release extraordinary amounts of energy. Tiny input. Massive output. And it made me wonder: What is the business equivalent of antimatter? Throughout history, competitive advantage has often come from discovering new forms of leverage. The steam engine allowed humans to multiply physical effort. Electricity transformed productivity. Computers amplified information processing. The internet dramatically reduced the cost of communication and distribution. Each shift allowed relatively small inputs to generate disproportionately large outcomes. Perhaps every era has its own version of antimatter. Not literally, of course. But economically. For a long time, retail’s leverage came from location. Then scale. Then supply chain excellence. Then digital reach. Today, I wonder if we are entering another transition. One where intelligence itself becomes the dominant source of leverage. Not intelligence in the abstract, but the ability to sense change, learn faster, and make better decisions than competitors. When I speak with retail leaders, very few struggle with a lack of data. Most are overwhelmed by it. The challenge increasingly isn’t collecting signals. It’s understanding which signals matter. Consumers continuously reveal preferences across marketplaces, search behavior, reviews, social platforms, and commerce ecosystems. Hidden within those signals are clues about what consumers may want next. The organizations that can interpret those signals faster may gain an advantage that is difficult to replicate. This is one reason I believe demand intelligence is becoming increasingly important. Not because it predicts the future perfectly. Nothing does. But because it may improve the quality and speed of decision-making across pricing, assortment, inventory, product development, and localization. Even a small improvement in those decisions can compound into significant business outcomes. Which brings me back to Musk’s tweet. Perhaps the most interesting part wasn’t antimatter. It was the idea that advanced civilizations may stop thinking primarily in terms of money and instead focus on more fundamental measures. Business may be heading through a similar shift. We will always care about revenue, profit, market share, and growth. But increasingly, I suspect the defining capability of successful organizations will be their ability to learn. How quickly can they detect change? How accurately can they interpret signals? How effectively can they convert insight into action? I don’t know if demand intelligence will become retail’s equivalent of antimatter. But I do think we’re entering an era where a small amount of insight can influence an enormous amount of value. And that’s a future worth paying attention to.

Result: 22% reduction in markdown dollars and 19% improvement in full-price sell-through within two seasons.

The shift was not in what products they carried. The shift was in how much capital each product received.