Shridhar Marri, CEO & Founder, Flyfish speaks about how his company harnesses the power of artificial intelligence (AI) to help e-commerce businesses increase sales, reduce customer acquisition costs and provide a transformative search experience.
The convenience of e-commerce is unparalleled. However, browsing through thousands of products is time-consuming and tiresome for consumers. For e-commerce businesses too maintaining catalogues, ensuring search yields accurate results is a constant struggle. Bengaluru-based Flyfish.ai is helping take the “pain out of search” for both consumers and online retailers by harnessing the power of Artificial intelligence (AI).
In the year and a half since its inception, Flyfish has worked with over 20 brands helping them increase their e-commerce conversions and sales by transforming their search capabilities.
The company has offices in Mumbai, Bengaluru, Gurugram, London, Toronto, New York, Chicago and Bellevue.
In an exclusive interview, Shridhar Marri, CEO & Founder, Flyfish speaks about how his company harnesses the power of artificial intelligence (AI) to help e-commerce businesses derive tangible benefits. Edited excerpts…
Tell us about what Flyfish does.
We help increase e-commerce sales and repeat purchases and reduce customer acquisition costs (CAC). That’s our reason for existence. How do we do that? By helping e-commerce businesses transform their search capabilities with artificial intelligence (AI).
Can you elaborate?
When people visit an e-commerce website, the search is their only interface with the business. But most online sites offer a traditional dumb search, based on keywords. But people’s needs are evolving. They want products or looks that they see on Instagram. But there are multiple products in a single look–not just a dress. There are shoes, accessories… We analyse the images they upload, understand all the products involved, search the client’s catalogue and fetch the products that match it or things that are similar to the client’s e-commerce site or app. So an AI-powered image search is the first.
The second is voice search…People today use a lot of voice search, especially on their mobile phones because it’s cumbersome to type. We provide a detailed voice search in over 100 national and international languages.
Tell us more about image search.
So customers can either upload an image from social media or their gallery or they can point the camera at something they like for instance a mannequin in a store window.
So all of your solutions are focused on search.
That’s one layer. The other layer is the AI data curator, which works on a client’s catalogue. Today, a lot of catalogues are in a mess–some do not have proper descriptions, and others do not have the right tags. Often, men’s products are tagged as women’s and so on. We have been through thousands of catalogues and identified the issues. When we come in, we sanitize the entire catalogue and curate it once and for all using machine vision algorithms.
Flyfish will look at all the images, and all products in a catalogue, and figure out the objects in each picture. Some images have four or five objects, our solution identifies and assigns the product type to each. Is it an accessory, a dress, a bottom or a top? After the product type is assigned, it goes further and decides product attributes. So if it’s jewellery, is it a pendant, if yes then is it gold or diamond? And it creates that database. Once the catalogue is revamped, most problems are solved because it does not ever need to be done manually again, except for inputting the price.
The third layer is guided selling, which simplifies the search. Let us say you go to Amazon today and search for a red kurta. It will throw up at least 10,000 plus products across hundreds of brands. Even after applying various filters, you are left with some 700, 800 products. You have to browse through all..read the descriptions, and reviews…to find the one that is right for you. This leads to decision fatigue. Now with AI, one can simply say “I’m going to a party, suggest something?” The AI chatbot which is part of the search will ask you some questions and then respond with text and images from within the the same website.
So instead of looking at 10,000 products, you can just look at a refined list of products suggested for you.
The three parts: smart search, AI curator and guided selling can be extended to all channels, whether it is WhatsApp, website, mobile app or Instagram.
Normally, companies have to look for multiple vendors to give each of these capabilities but we provide it all in a single place.
When does the ROI start for a brand working with you?
We generally take about four to six weeks to go live with any of the features. Things on your brand right now. After that, brands can see the impact in about four to six weeks. First, they will notice that their search volume and conversions will increase.
How do you measure impact?
Apart from search volume and conversions, the metric we measure is average order value (AOV), which also increases because you are engaging differently and you are helping customers discover the right products, right away. When the search is time-consuming and cumbersome, it leads to a lot of drop offs, which reduce because the research time is reduced by almost 80% and the overall buying cycle is much faster. And a better experience leads to customers coming back, so we also consider repeat purchases.
So your product is ideal for all categories?
We work across the broad spectrum. Right now we’re working on fashion, cosmetics, jewellery, consumer electronics and furniture among others. We have been approached by several large banks as well.
Similarly, some real estate players have approached us as well to help with the massive list of leads.
So we have built a property advisor for one of the property developers in Mumbai. It answers questions and helps narrow down the search. It is increasing the leads by 20%.
What is next for Flyfish?
We want to bring a world-class customer experience for Indian brands. In retail, today almost in 90% of the cases, the searches are suboptimal and brands do not have a solution for this right now. Even the largest of the large existing search companies are all legacy technology companies. They all use keywords, which produce inaccurate results if the tagging is wrong and incomplete. A lot of manual work is required, which is a pain. We want to remove the pain for retail from a search point of view. Search should work like a beautiful experience.
Brands like Amazon have invested millions of dollars in perfecting their search engines but not all companies have that kind of financial muscle. We want to be that bridge and help all retailers at a price point they can afford to use technology which matches the world-class. And we want to cover the entire retail segment, including all subverticals.
