The AI Blueprint For Fashion Ecommerce

Fashion ecommerce is entering a new era. The first was about adding intelligence to the customer journey: better search, smarter recommendations and chatbots that could answer routine or generic queries. The current phase is more dynamic, with AI reshaping marketplace operations. 

Instead of asking shoppers to browse thousands of products, marketplaces are trying to build experiences around each individual. Search, recommendations, styling, fit and even storefronts are becoming more dynamic than ever.

Our conversations with Myntra’s leadership, a look at Nykaa’s latest investor roadmap, and Meesho’s own AI rollout point to this shift. At Myntra’s Tech Day, CPO Lakshminarayan Swaminathan described this as reimagining the shopping journey. 

In practice, this means:

• Personalised search tailored to each shopper, rather than broad segments such as cities or demographics
• Homepages evolving into algorithmic storefronts, unique to every customer
• AI-generated styling and outfit visualisation, with size and fit tools beginning to simulate how garments look across different body types

The same is true for conversational commerce. Myntra’s shopping assistant, ‘Maya’, is an early step in this direction. According to CTO Pramod Adiddam, the future is not about having a chatbot on top of every application but making every experience conversational. In other words, conversation is the new operating layer between shoppers and marketplaces.

“Myntra is careful not to deploy AI just because it exists,” Adiddam said, adding that every customer-facing capability starts with a business hypothesis, tested on a small group of users before a full rollout. 

Today, 85% of Myntra users with sufficient shopping history get AI-powered size & fit recommendations, with future versions expected to visualise garments on the customers themselves.

Nykaa is chasing something similar. The company is building Ask Nykaa, its virtual closet and personalised storefronts on first-party behavioural data.

Meanwhile, Meesho is doing its own version of voice AI with Vaani, a GenAI shopping assistant that handles product discovery, comparisons, payments and checkout in Hindi and English. Over 1.5 Mn users interacted with it within a month of launch.

Put together, these moves suggest the next phase of ecommerce may be less about helping people search faster, and more about making search unnecessary altogether.

Fashion Ecommerce Engine Gets An Upgrade

While AI is most visible in customer-facing features, its bigger impact is behind the scenes. A case in point is Myntra, which has rolled this out across catalogue creation, merchant onboarding, customer support, supply-chain planning and internal analytics, cutting down manual work at every step. 

The results:

• Merchant onboarding is now down to 2 days from 15 days 
• Product listing now takes only 3-4 hours
• 31% of customer support queries are now automated
• Customer satisfaction has gone up 2X
• Supply-chain simulations reduced to just 1 hour for 2 days
• Business analytics is now 10X faster, powered by its internal platform, BIRA

The bigger game, however, is the infrastructure behind these. Rather than building isolated AI applications for every department, Myntra has created reusable internal platforms that multiple teams can build upon. 

Nykaa appears to be following a similar path. The company now positions its One Nykaa AI Platform as the shared infrastructure connecting customer experiences, enterprise workflows and employee productivity, with more than 55% of its FY26 technology impact attributed to AI initiatives.

Meesho has also open-sourced BharatMLStack, its end-to-end machine learning infrastructure that was previously used internally. The stack allows startups to build, deploy, and scale AI applications more cost-effectively.

Infrastructure Becomes The New Battleground

If the first generation of ecommerce companies competed on assortment, pricing and logistics, the next may compete on infrastructure. Myntra avoids betting on a single model provider. Its size & fit engine is built in-house, while recommendation explainability leans on frontier APIs, and everything in between draws on whatever mix works best. Swaminathan says the goal is to stay flexible, since the landscape shifts too fast to lock in.

As models themselves become commodities, the real edge is shifting to proprietary data, evaluation frameworks and platforms that let teams ship faster. Nykaa calls its One Nykaa AI Platform its ‘true moat’, built on first-party data and shared infrastructure. Myntra is doing something similar, turning successful experiments into reusable platforms across search, voice, analytics and engineering.

That is probably where Indian fashion commerce is headed. The race is no longer about AI for its own sake, but about redesigning how marketplaces run underneath. Customers won’t see any of this. They’ll just notice that recommendations feel sharper, sellers come online faster, and decisions that took days now take minutes.

Top Stories From India & Around The World

• Vorflux Lands $15 Mn: Former Rippling cofounder Prasanna Sankar has raised $15 Mn for AI software engineering startup Vorflux, backed by Y Combinator, Peak XV Partners and others. The investment underscores growing demand for AI platforms that automate the entire software development lifecycle using specialised AI agents rather than standalone coding assistants.
• Reo.Dev Raises $11.3 Mn: AI sales intelligence startup Reo.Dev has secured $11.3 Mn in a Series A round led by Elevation Capital to accelerate product development and deepen its US presence. The raise signals continued investor appetite for Indian AI startups building enterprise software for global markets, particularly developer-focused sales and go-to-market platforms.
• NVIDIA Unveils GPU Financing Model: NVIDIA has introduced a revenue-sharing financing programme for AI cloud providers, allowing them to procure GPU infrastructure with lower upfront capital requirements. The move could accelerate GPU deployment across Indian neoclouds such as Yotta, NxtGen, Neysa, NeevCloud and E2E Networks, while strengthening NVIDIA’s influence over the country’s AI infrastructure.
• Cabinet Clears ₹1.28 Lakh Cr Semicon 2.0: The Union Cabinet has approved the ₹1.28 Lakh Cr Semicon 2.0 programme to boost India’s semiconductor ecosystem across chip design, manufacturing, packaging, and R&D. The move reinforces India’s push to reduce import dependence and emerge as a global semiconductor hub.
• Nuclear Plant Files Surface On Dark Web: Nearly 19,000 files allegedly linked to the Kudankulam Nuclear Power Plant have surfaced on the dark web following a ransomware attack involving contractor Reliance Infrastructure. The incident highlights the growing cybersecurity risks facing India’s critical infrastructure and supply chains as CERT-In investigates the breach.

The Weekly Buzz: Moonshot’s Kimi K3 Triggers The Open AI Race

Moonshot AI has unveiled Kimi K3, a 2.8-Tn-parameter open-weight model that delivers frontier-level performance across coding, reasoning and agentic AI tasks, while offering native multimodality and a 1-Mn-token context window. The launch sparked overwhelming demand, forcing the company to temporarily pause new subscriptions as it scales infrastructure ahead of a broader open release and its planned IPO.

The launch coincided with the World AI Conference (WAIC) in Shanghai, reinforcing China’s accelerating push in open-weight AI just as competition with US frontier labs intensifies. Kimi K3 has also reignited the debate around open versus closed AI models, with developers increasingly gravitating towards highly capable models that can be customised and deployed at lower costs rather than relying exclusively on proprietary offerings.

The bigger question now is execution. Running inference for a model of this scale without compromising latency, reliability or economics will be a significant challenge, especially as adoption surges. But irrespective of how Moonshot manages that transition, Kimi K3 marks another milestone in the shift towards open-weight AI. Frontier performance is no longer confined to closed ecosystems, forcing the industry’s biggest players to rethink how they compete on accessibility, pricing and developer adoption.

Startup In The Spotlight: Linkrunner  

Founded in 2025 by Shreyans Sancheti and Darshil Rathod, Linkrunner is building an AI-powered mobile measurement platform that helps app businesses accurately attribute installs, track user journeys and measure the effectiveness of their marketing campaigns from a single dashboard.

Instead of relying on fragmented attribution and analytics tools, the startup combines attribution, deep linking, cohort analysis, retention tracking, revenue measurement, remarketing, anomaly detection and creative performance insights into one platform. It aims to help marketers understand how users discover, install and engage with their apps while optimising campaign performance and marketing spend.

Born out of the founders’ experience building internship platform Bluelearn, Linkrunner has set out to create a simpler and more affordable alternative to existing mobile attribution platforms. Today, it follows a usage-based SaaS model and serves more than 150 customers across 15 countries, including India, the US, Europe, Southeast Asia and the UAE. The startup claims to have processed over 6 Bn data points and now analyses nearly 1 Bn data points every month.

Looking ahead, Linkrunner is betting that as mobile user acquisition becomes more expensive and privacy regulations continue to reshape digital advertising, businesses will increasingly need AI-powered measurement platforms that deliver more accurate attribution, deeper customer insights and better visibility into marketing ROI.

Prompt Of The Week

What prompts and hacks are CTOs, CEOs and cofounders using these days to streamline their work? 

Here’s the prompt Subhash Kalluri, founder of FreJun, uses to make AI act as a second reviewer, identifying mistakes, inconsistencies and improvement opportunities before a piece of work is finalised.

“You are a world-leading expert in B2B SaaS AI-led growth hacking.

I need you to:

• Audit the answer in a loop, checking for biases, blind spots, unstated assumptions, overestimations, and factual mistakes
• Correct whatever the audit surfaces, then run the pass again, and keep going until no new issues come up
• Flag anything you are still uncertain about instead of smoothing over it
  
• List every assumption the final answer rests on, set out clearly alongside the output”
  

Editor’s Note: Some prompts may need to be adjusted by users for best results or may not work as intended for certain users.

[Edited by: Shishir Parasher]
[Creatives by Abhyam Gusai]

The post The AI Blueprint For Fashion Ecommerce appeared first on Inc42 Media.