Prabhanu Kumar Das, Medianama
16 July 2025

E-commerce logistics platform Shiprocket announced the launch of Shunya.ai, a sovereign AI model developed in India to support the country’s Micro, Small and Medium Enterprises (MSMEs), on July 11. The company claims that it is India’s first multimodal AI stack, built in partnership with US-based Ultrasafe Inc.

This announcement comes at the heels of Shiprocket filing a confidential draft red herring prospectus (DHRP) with the Securities and Exchange Board of India (SEBI) in May 2025 for their Initial Public Offering (IPO). The company is expected to raise around Rs 2500 crore in its IPO.
What does the AI model offer?

As per Shiprocket’s website, Shunya.ai is built on a freemium model, with unlimited access priced at Rs 499 a month for MSMEs. It is directly integrated into the Shiprocket platform and offers AI agents across multiple languages. According to the company, the agents can perform the following tasks:

1. Catalogue management and creation: It automates the creation and management of catalogues, and enables product listings in multiple languages.
2. Ad campaign creation: It can assist in generating marketing campaigns in multiple languages as well as in creating the advertising content.
3. Automated customer support: Offers AI chatbots for customer support.
4. Streamlining delivery and logistics: The model can find the most efficient and affordable methods for delivery, as well as tracking orders.

Shiprocket CEO Saahil Goel stated, “We’ve adapted Shunya.ai from the ground up for Indian languages, commerce workflows, and MSME needs. By embedding it directly into our platform, we’re giving over 1,50,000 sellers instant access to tools that are intelligent, local, and scalable, levelling the playing field for businesses across Bharat.” Notably, Larsen and Toubro’s AI cloud arm, Cloudfiniti is reportedly providing the underlying GPU infrastructure, ensuring that all data processing and storage remains within India.

This AI model does offer multiple benefits but it will not level the playing field against big players, as per Devangshu Dutta who is the founder of specialist consulting firm, Third Eyesight.

“While Shunya AI can help small businesses compete better, it won’t completely level the playing field. Large companies still have greater organisational capacity and capability to respond to the insights offered, including more data and bigger budgets. The real benefit for small businesses is improving how they work and serve customers within their current markets, rather than suddenly competing with giants,” Dutta said.

The E-Commerce AI Pivot

This is not the first time that an Indian e-commerce platform has unveiled a B2B AI service through its existing platform. Zepto recently launched Zepto Atom in May 2025, a real-time tool that offers consumer brands available on the platform minute-level updates, PIN-code level performance maps, and Zepto GPT, a Natural Language Processing (NLP) assistant trained on internal data that brands can query about their stock keeping units (SKUs) and performance data.

Zomato and its e-commerce arm Blinkit have also been growing their AI capabilities. Analytics India Magazine previously reported that the company’s generative AI team has grown from 3 to 20 engineers in the time-span of a year. Zomato introduced a personalised AI food assistant for users, and also uses AI in its backend to optimise delivery times and improve consumer support. Blinkit also released the Recipe Rover AI in May 2023, an AI assistant for recipes.

Other companies like Swiggy with ‘What to Eat’ AI, Myntra’s MyFashionGPT AI shopping assistant, and Amazon’s Rufus have also adopted AI assistants on their platform as a tool for the consumer.

The issue of merchant stickiness

Dutta asserts that this shift means platforms like Zepto and Shiprocket are changing from being service providers to becoming data intelligence companies. They are generating, or are in the process of generating revenue through transactional data that flows through the company.

“While this can create better insights and automation for merchants on these platforms, it also could make the merchants more dependent on the platforms. Once a merchant builds its operations around a platform’s specific AI tools and insights, it becomes much harder to switch to a competitor – creating stronger merchant stickiness. We already see this in infrastructure and core services such as banking and financial services, enterprise cloud services, building management etc. and the same is likely to happen in AI-enabled process management”, he said.

Why this matters

As Shiprocket is preparing for an IPO, Shunya.ai becomes another means to generate revenue for the company. This app can extend Shiprocket’s reach to local physical stores and MSMEs, by offering them the opportunity to provide the same experiences and support to the consumer that larger retailers and e-commerce platforms do, while automating delivery automation, cataloguing, and customer support.

Furthermore, the launch of this model is also part of the larger trend of AI integration and automation, both within e-commerce platforms for their consumers and within the back-end for optimisation.

Competition in these sectors and merchant stickiness may also become an issue, as businesses hosted on these e-commerce services may become reliant on specific AI tools and their outputs.

Questions of data privacy are also important when it comes to service companies moving towards data intelligence: How do these AI models gather and use data? The consent of end-consumers in these B2B models, data storage, and security are all issues that need to be studied as e-commerce and retails pivots towards AI.

Some Unanswered Questions

MediaNama has reached out to Shiprocket with the following questions and will update the article when we receive a response.

1. How does Shunya AI differentiate itself from other global or domestic AI tools being used in the logistics and e-commerce sectors such as Zepto Atom or Shopify Magic?
2. What data is Shunya AI trained on? Is the training dataset sourced exclusively from Shiprocket’s operations, or are third-party data streams also used?
3. What data will Shunya AI’s marketing campaign models access? How will it ensure privacy and data protection of the end consumer of the business who is using these models?
4. How does Shiprocket ensure compliance with Indian data protection laws, especially given the scale of customer and seller data being used?

(Published in Medianama)

The apparel retail sector worldwide thrives on change, on account of fashion as well as season.

In India, for most of the country, weather changes are less extreme, so seasonal change is not a major driver of changeover of wardrobe. Also, more modest incomes reduce the customer’s willingness to buy new clothes frequently.

We believe pricing remains a critical challenge and a barrier to growth. About 5 years ago, Third Eyesight had evaluated the pricing of various brands in the context of the average incomes of their stated target customer group. For a like-to-like comparison with average pricing in Europe, we came to the conclusion that branded merchandise in India should be priced 30-50% lower than it was currently. And this is true not just of international brands that are present in India, but Indian-based companies as well. (In fact, most international brands end up targeting a customer segment in India that is more premium than they would in their home markets.)

Of course, with growing incomes and increasing exposure to fashion trends promoted through various media, larger numbers of Indian consumers are opting to buy more, and more frequently as well. But one only has to look at the share of marked-down product, promotions and end-of-season sales to know that the Indian consumer, by and large, believes that the in-season product is overpriced.

Brands that overestimate the growth possibilities add to the problem by over-ordering – these unjustified expectations are littered across the stores at the end of each season, with big red “Sale” and “Discounted” signs. When it comes to a game of nerves, the Indian consumer has a far stronger ability to hold on to her wallet, than a brand’s ability to hold on to the price line. Most consumers are quite prepared to wait a few extra weeks, rather than buying the product as soon as it hits the shelf.

Part of the problem, at the brands’ end, could be some inflexible costs. The three big productivity issues, in my mind, are: real estate, people and advertising.

Indian retail real estate is definitely among the most expensive in the world, when viewed in the context of sales that can be expected per square foot. Similarly, sales per employee rupee could also be vastly better than they are currently. And lastly, many Indian apparel brands could possibly do better to reallocate at least part of their advertising budget to developing better product and training their sales staff; no amount of loud celebrity endorsement can compensate for disinterested automatons showing bad products at the store.

Technology can certainly be leveraged better at every step of the operation, from design through supply chain, from planogram and merchandise planning to post-sale analytics.

Also, some of the more “modern” operations are, unfortunately, modelled on business processes and merchandise calendars that are more suited to the western retail environment of the 1980s than on best-practice as needed in the Indian retail environment of 2011! The “organised” apparel brands are weighed down by too many reviews, too many batch processes, too little merchant entrepreneurship. There is far too much time and resource wasted at each stage. Decisions are deliberately bottle-necked, under the label of “organisation” and “process-orientation”. The excitement is taken out of fashion; products become “normalised”, safe, boring which the consumer doesn’t really want! Shipments get delayed, missing the peaks of the season. And added cost ends in a price which the customer doesn’t want to pay.

The Indian apparel industry certainly needs a transformation.

Whether this will happen through a rapid shakedown or a more gradual process over the next 10-15 years, whether it will be driven by large international multi-brand retailers when they are allowed to invest directly in the country or by domestic companies, I do believe the industry will see significant shifts in the coming years.

It has been around 200 years since the birth of Charles Darwin, and about 150 years since the publication of his and Alfred Wallace’s thoughts on evolution by natural selection. In their honour, let us remind ourselves of the basic theory that all of us learn at school. (So I’m a few months late acknowledging it – please bear with me!)

On Evolution: Change Happens

(1) Species differ from each other, but individuals within a species also differ from each other quite a bit.

(2) These differences are due to changes to the basic genetic framework of the organism (mutations) which can get passed on to following generations.

(3) The environment keeps changing physically, climatically and biologically.

(4) In the new (changed) environment some of the mutations survive better than others (“natural selection”).

(5) The effect of these changes over several generation results in the evolution of species, and the rise of new species.

The primary reason I am highlighting this theory is because, to my mind, businesses are like living beings. Businesses are conceived, given birth to, they grow, and most of them die after a few years or a few decades. During their life some businesses get married (merged or acquired), and sometimes they give birth to other businesses.

About 2-3 years ago, the business climate seemed predictable and only looking upwards – the biggest challenges in the food and grocery sector seemed to be whether your ambition was bigger than your competitor’s. Many predictions were made about how the large – more “organised” – businesses would quickly kill the small.

However, with much turmoil in the business environment in the last year or so, it is evident now that it is not just the small companies that are vulnerable. The change in the environment is also giving new growth opportunities to the smaller or younger, previously vulnerable, businesses. While some of the larger businesses have died or are in the process of dying, some of the smaller businesses are mutating even more to survive better in the changed surroundings.

Although small businesses are always looking for growth, the new environment can bring such a surplus of opportunities that, in the helter-skelter growth the learnings are quickly lost and the business may actually go off the tracks.

On Process: Passing On the Genes

The challenge for the smaller businesses now is to pass on their genes down the generations; for the management to ensure that the newer stores and the newer recruits gain from the learning and the adaptations already in the organisation.

At an entrepreneurial stage, the core team handles critical activities and is on call to guide others. The team is knit quite tightly, and located geographically close together. The stores are few and in locations with a similar environment. “Knowledge” is inherent in the way you do things, guided rather than taught.

You may recall my stressing culture and organisational personality, the “people” end, in a previous article. At the early stage of the business, very often, that is all there is. But growth needs replication and predictability.

Biology again gives us a great lesson in how to replicate learnings and functionality: genes (DNA) provide the template for cell functions, and are reproduced almost faithfully from previous generations.

In a business, such replication comes from well-designed processes incorporating the intent, the activities and the desired outcomes. For growth, processes are a must; they are the genetic code of the business. Processes provide the design for how a customer would interact with the store, how the store would interact within itself and with other points in the organisation, and how the organisation would interact with external agencies.

You may ask, “How much process should we depend on, and how prescriptive or restrictive should we make them?” You may also point out that processes start off with very good intention, but with time – and often distance from head office – the processes decay.

And you would be right.

On Decay: Bad or Good

Even in bureaucratic organisations, adjustments are made to fit people or situations, and that causes the process to mutate.  Sometimes the change is temporary, at other times the process may change completely and permanently. If changes happen passively and are not channelled the existing process will decay.

I use the word “decay” carefully. While the process change itself may be good at a point (e.g. responding to a customer need), the organisation as a whole may not learn much from it, or the change may affect one part of the organisation and not others. If that happens, the organisation and its systems will become dysfunctional at some point.

For instance, it could be the little leeway that the merchandising head provided to some managers that erupts into an uncontrolled working capital epidemic across the chain. Or a margin adjustment with a vendor at a certain point in time becomes a deadly norm.

So, back to evolution: mutations are a fact of life. Adaptations are happening because of the changes in the environment. Managers need to critically question: does this change meet a current ongoing need or provide an ongoing advantage, and can it apply to the rest of the business? If the answer is no, ask people to read the rule-book (the process manual).

If the answer is yes to both, change the rules, and make sure the new process is implemented quickly and smoothly across the organisation.  Then it will be “adaptation” rather than “decay”.

After all, the conclusion that Darwin, Wallace and many others have given us is this: it is not the strongest, the biggest, the fastest, but the most adaptive who survive.

It often seems like “Scaling Up = Dumbing Down.”

As an organisation starts to scale up, the challenge is getting everyone to work the same way as earlier. And Process with a capital “P” takes over.

In the absence of inspiring and totally involved leadership, it often is just a hop and a skip from Bureaucracy.

If a lot of people feel the same way, there may yet be hope for small businesses and the more personalized touch.