AI changes this now
AI does not just improve search. It compresses decision-making from days to seconds — and makes it personal.
But capability alone is not enough. The real shift is behavioural. 50 million shopping queries are already happening on ChatGPT every day. 64% of Indian consumers are now using GenAI for shopping decisions. Users are comfortable asking before they buy. Decision-making is being decoupled from marketplaces. AI-referred shoppers convert 31% better than standard traffic, and are 68% less likely to return what they bought.
For the first time, the decision layer can exist as its own product. Google built a $4 billion annual revenue business in India by owning search intent. Amazon and Flipkart built $1.2 billion by owning product discovery. The decision layer sits upstream of both — and it is still unclaimed.
Why this will not be owned by general AI or marketplaces
The obvious question is why ChatGPT or Perplexity cannot own this layer. The answer is not capability. It is structural misalignment. A true decision layer requires four things that neither general models nor marketplaces can build.
Shopping is visual and comparative. Decisions require product cards, side-by-side comparisons, and category-specific interfaces — not paragraphs of text. Flash is built to know every option in a category and surface what is genuinely best, regardless of who paid to be visible.
Not just top results — every product, price, and merchant, continuously updated. General models have no real-time commerce data and no incentive to build it. They cannot tell you what a product costs today, or whether it was cheaper last week.
Preferences, purchase history, context. General models start from scratch every session. Flash builds your profile over time and uses it to anticipate what you need — not just respond to what you asked.
Marketplaces cannot recommend competitors. Amazon cannot tell you to buy on Flipkart. Flipkart cannot tell you a competing product is better. They generate billions from sponsored listings — they are structurally unable to be unbiased.
What Flash is building
Every person deserves the truth about what they are buying. Flash is how they get it.
Flash is the layer between intent and purchase. Before you buy anything that matters, you ask Flash. In seconds, Flash tells you: is it actually good, is it right for you, and where should you buy it — across every product, every category, every merchant. No login. No friction. No bias.
Built on commerce-specific data that general models do not have and cannot quickly replicate, with a seven-month head start on building it. Every interaction improves the system. Every decision makes the next one better. This is not a feature. It is a compounding advantage.
Flash grows in three stages, each building on the last.
The core research product — synthesising expert reviews, YouTube, Reddit, and verified buyer signals into a clear, structured verdict, accessible instantly from any surface. Zero friction, no login required. The goal is trust: the kind that no marketplace can build because they are structurally conflicted, and no general model can build because they do not have the data depth.
Skin analysers, face shape analysers, ingredient compatibility engines, spec-weighted scoring for electronics. Personalised storefronts where Flash knows your preferences and surfaces what is right for you before you ask. Flash stops being a tool you reach for and becomes the layer you rely on. Brands pay to be present where purchase decisions are made — the B2B surface that makes Flash indispensable to every brand in Indian commerce, not just every shopper.
Today a buy-now click exits to Flipkart or Nykaa. Tomorrow it completes inside Flash. Express checkout, agentic reorder for repeat purchases, deep merchant integrations. Flash moves from advising decisions to making them. The full commerce loop — research to purchase — on one surface.
Flash is also building directly for brands. As AI reshapes how shoppers research and decide, brands need to be present in that layer too. Flash offers AI shopping agents that supercharge brand storefronts with conversational product guidance and personalised recommendations (Buffy AI), and intelligence tools that help brands understand how they are being discovered across AI platforms (Pulse AI). Flash sits on both sides of the decision layer — consumer and brand.
How people access Flash
We do not ask users to change behaviour. Every Flash user arrives at a product page with intent — they are already considering a purchase. Flash sits at that exact moment, before the decision is made.
That is not just a distribution mechanic. It is a structural position. By intercepting intent at scale — across every product, every merchant, every surface — Flash has the ability to influence which product a shopper chooses, and where they buy it. That is demand-shaping power. It compounds as the user base grows, and it costs almost nothing to acquire.
Prepend flash.co/ before any product URL on any marketplace. Full research view instantly — no account, no install, no friction.
flash.co/amazon.in/dp/B08…Share any product from Amazon, Flipkart, Myntra or any browser directly to Flash. Native to how Indian shoppers browse on mobile.
iOS & Android appInstalled once, activates automatically on every product page across any e-commerce site. Research appears without switching tabs.
Available on Chrome Web StoreThe numbers so far
We launched in September 2025. Seven months in, here is an account – the metrics that matter, and the ones that still need work.
Sep '25 – Mar '26
to date
(~$0.05)
user growth
| Metric | Sep '25 | Oct '25 | Nov '25 | Dec '25 | Jan '26 | Feb '26 | Mar '26 |
|---|---|---|---|---|---|---|---|
| Monthly Users | 111k | 200k | 306k | 505k | 565k | 933k | 1.04M |
| Products Researched | 207k | 358k | 588k | 1.06M | 1.32M | 2.07M | 2.6M |
| CAC (₹) | ₹17.5 | ₹10.1 | ₹8.2 | ₹4.3 | ₹3.6 | ₹2.2 | ₹2.1 |
| Tech cost / product | ₹12.4 | ₹3.8 | ₹2.4 | ₹1.9 | ₹2.6 | ₹1.7 | ₹0.9 |
| # | Platform | Country | Monthly Visits | Avg Duration | Pages/Visit | Bounce Rate |
|---|---|---|---|---|---|---|
| 1 | Flash AI | India | 2.6M | 3:36 | 5.99 | 30.9% |
| 2 | Dupe | US | 975k | 1:47 | 4.18 | 37.0% |
| 3 | StyleDNA | US | 424k | 1:26 | 4.35 | 46.6% |
| 4 | Phia | US | 188k | 1:24 | 2.73 | 41.7% |
| 5 | Alta | US | 178k | 2:09 | 4.05 | 51.1% |
| 6 | Vetted | US | 42k | 0:51 | 1.79 | 57.6% |
| 7 | Daydream | US | 39k | 0:52 | 3.38 | 42.2% |
Who uses Flash – and why it matters
What we are building next
The long-term roadmap is set. The next 90 days are about three things: making Flash the discovery destination, building the habits that bring users back, and converting research into revenue.
Ask Flash what to buy in plain language. Share a reel or a screenshot and get instant research on the products in it. Scan a shelf in-store and compare alternatives in real time.
Skin analysers, ingredient compatibility engines, face shape analysers for BPC. Structured spec-comparison interfaces and shelf-scan research for electronics and appliances. The features that make Flash irreplaceable in a category — and that give users a reason to open Flash before they even have a product in mind.
Three levers in parallel: high-intent advertising on product research pages, AI shopping assistant deployments on brand storefronts, and in-session checkout that closes the purchase without leaving Flash. Each earns independently. All three are in motion.
How we think this will scale
Three revenue streams: affiliate commissions, advertising inventory, and agentic checkout. All three need scale. We are building for 10× the scale by end of 2027.
Affiliate marketing in the US is a $13.6 billion industry. It accounts for 9.4% of all US e-commerce sales and grew 49.8% from 2021 to 2024. India's affiliate market reached $331 million in 2023 and is projected to exceed $420 million in 2025. India is roughly 1/30th the size of the US market. That gap is closing.
India now has the conditions for affiliate economics to work. E-commerce GMV is $60 billion in 2024, forecast to reach $170–190 billion by 2030. Quick commerce reached $6–7 billion in 2024. Multiple marketplaces and direct-to-consumer channels now compete for the same shopper. D2C alone has 11,000+ active brands, with roughly 800 funded and operating at scale. Enough merchant competition, enough purchase intent, enough SKU depth.
Commerce advertising in India is already large. Amazon India and Flipkart combined for ₹14,652 crore in ad revenue in FY25. Retail media is 22% of India's digital ad spend, growing at 40% CAGR. Flash captures shoppers at purchase intent. That is tier-1 advertising inventory.
The economics work in three stages — each earning independently, each compounding with scale.
5–15% on completed sales. Beauty and personal care brands pay up to 22.5% — among the highest affiliate rates in Indian e-commerce.
Shoppers arrive on Flash with a product in mind and a decision to make. That is the highest-intent surface in commerce — and it commands premium rates.
The model shifts from click referrals to completed purchases. Commission rates on completed transactions are 3–4× higher than click-through models.
At 10× current scale by December 2027, Flash enters a $10 million ARR horizon. Beyond that, revenue scales disproportionately with user growth.
Flash intercepts purchase intent at the moment it is highest — when a shopper has a product in mind and is about to decide. Every session is a chance to shape that decision: which product, which variant, which merchant. The data already shows what that looks like at scale: inbound traffic from D2C and other platforms is 18.3%; outbound clicks to those same merchants are 34.9%. Amazon's share shifts from 40.5% inbound to 25.1% outbound. Flipkart from 29.2% to 15.5%. At 3.5 million users, Flash is already reshaping where India buys.
The belief
Shopping doesn't need more choice. It needs clarity.
Google helps you find.
Marketplaces help you buy.
Flash helps you decide.
Once the decision layer is owned, the rest of commerce follows.