This startup by former Mindtree employees uses an AI platform to up the game of online shopping

AskSid is a Bengaluru-based startup that helps online brands train algorithms on their product data. Its AI bot also assists online buyers with easy product discovery and answers all their questions.

14th May 2019
  • +0
Share on
close
  • +0
Share on
close
Share on
close

Dolly Roy, an expectant mother working in the Netherlands, was looking for a pair of warm leggings. Logging on to some ecommerce websites, she was overwhelmed by the number of choices staring back at her. There were over 400 options of warm leggings, and Dolly was at sea. Confused, she decided to go to the nearby store and get a salesgirl's help. Finally, she purchased a pair of leggings. 


Dolly is not alone in her confusion. With a multitude of choices, irrespective of where you are, shopping can become fairly stressful. The decision fatigue and choice paralysis that arises from the problem of plenty had Dolly’s husband, Sanjoy Roy, thinking. He saw there was a clear knowledge gap between the end user and the product.


shopping, ecommerce, AI, artificial intelligence, AskSid

Choice paralysis: Shopping has become a stressful activity given the rise of a plethora of brands vying in a crowded market.

(Image: Artem Beliaikin on Unsplash)

“But the brands that make the products have answers to all possible questions. That is what led me to start AskSid.ai, a conversational AI platform,” says Sanjoy. 


Back in India, Sanjoy started the platform with Dinesh Sharma, his colleague from Mindtree


The specialised conversational artificial intelligence (AI) platform assists online buyers with product discovery, answers queries, raises customer service requests, and loops in human agents to answer them. All this across channels, and in multiple languages. 


The company claims an Austria-based women’s apparel brand, which is Europe’s largest, as its first major client.


How AskSid works 


AskSid mainly delivers three business outcomes – a) conversations with shoppers to simplifying their experience, b) a better product catalogue to showcase products to the customer, and c) marketing insights for companies.


To this end, the AI platform offers a host of components: 


  • AskSid: The company’s flagship conversational AI platform 
  • Replay-Sid: Allows brands to track consumer conversations 
  • Switch-Sid: Allows the handover from the AI platform to a human agent 
  • Measure-Sid: the analytics component to generate insights
  • Monitor-Sid: A business dashboard give brands business KPIs in real time 
  • Intelli-Sid: Hosts the AI domain metadata 
  • Train-Sid: Cognitive workbench for data enrichment, and algorithm testing 
  • AskSid Canonical Data Store: Raw data from a brand is used by algorithms for data enrichment and other functions.


“Our platform comes with its own canonical data model for retail, and the raw data from a company is first ingested to this data store, and algorithms use this raw data for data enrichment, model training, model testing, etc,” Sanjoy adds.


shopping, ecommerce, AI, artificial intelligence, AskSid

Sanjoy Roy and Dinesh Sharma, Founders of AskSid


Also read: Conversational AI will change the way people interact with devices



Competition


Conversational AI platforms are not new. The first wave started in 2014-15 as several companies and brands built their own chatbots. Prominent players in the space include Mad Street Den’s Vue.AI, which recently raised $17 million in Series B funding. Using computer vision and AI, Vue.ai focusses on building an end-to-end stack that helps retailers achieve scale. 


“We, however, are a vertical-focused AI offering deep focus in retail and consumer goods. We have our own ‘intents’ library and ‘NER’ (Named Entity Recognition) model for retail, and more and more conversation mean our intent and entity library are growing all the time,” says Sanjoy. 



Also read: Will the real AI startups please stand up?



What sets it apart 


While the team did not disclose the pricing, AskSid follows a SaaS-based enterprise model, with a monthly subscription fee. The team claims that having a large apparel company as its first client, it has identified more than 75 unique shopper intents, and that its model can accurately predict the intent against every incoming user message.


On AskSid’s first client, Sanjoy says, “The first go-live with this brand happened in December 2017 for their Ireland market, and then we rolled it out to 14 other countries by the middle of 2018. In early 2019, we launched our solution to this brand’s Facebook page and Skype channel too.” 


The team currently claims to have over 25 different installations of their product.


“We can now configure ‘next best action’ for queries and conversations with customers and this allows us to support multi-turn conversations, something pure play chatbot companies cannot offer,” he adds. 


Numbers and future 


While most customer service-oriented AI platforms attempt to address FAQs and general information, Sanjoy says AskSid is going deep to continuously enrich information on the product catalogue. 


“The outcome AskSid produced was a staggering 40 percent increase in conversions compared to web channels and automation of 800 customer service requests with creation of 1300+ size profiles,” says Sanjoy. 


Currently bootstrapped, the team raised initial angel funding led by Krishnakumar Natarajan (NKK), Chairman, Mindtree, and Rajan Anandan, ex-MD, Google SE Asia, and currently at Sequoia Capital. 


“With the base product built out to a large extent, our plan is to now scale up in terms of new customer onboardings, and also going deep in the vertical domain for retail and consumer goods (CG). In the end, we want to be known as the best conversational AI for retail and CG,” says Sanjoy. 



Also watch: Artificial intelligence will change everything – marriage, sex, work, says Manthan’s Atul Jalan


  • Facebook Icon
  • Twitter Icon
  • LinkedIn Icon
  • WhatsApp Icon
  • Facebook Icon
  • Twitter Icon
  • LinkedIn Icon
  • WhatsApp Icon
  • Share on
    close
    Report an issue
    Authors

    Related Tags