Artificial Intelligence (AI) e-Commerce retail is huge and right now and we haven’t yet scratched the surface. “There is a lot to do,” says Sudhir Kadam, Venture Partner of FYDA Growth Partners at the 7th annual Big Data Conference held last month with the theme, “Al.”
Referring to big players in e-Commerce, Kadam said, “Even if you see Amazon and Walmart, [they are] still searching product categories, searching [for] your products should be auto-categorized.
Kadam, who moderated the keynote panel: AI in Retail/E-Commerce, which had speakers from GAP, Levi Strauss, Walmart Labs, Salesforce, and Aible said, “The challenge is [that] it’s hard to dismantle a legacy system for big companies, though it’s easy to adopt if you are starting from scratch.”
Matthew Tovbin, Software Architect at Salesforce, commented on AI, “It’s simple, it’s transforming [from the] inside out.”
He said that companies like Salesforce have to innovate and we embraced AI few years ago and we are truly integrated.
“Customers come to us with more request for more features and more stuff,” Tovbin said. “So people have to embrace [that] and as a software company your engineers constantly learn new stuff.”
Taran Singh, Director at GAP, says at Gap, in general, they are at the forefront of understanding where IA use cases are relevant both from a scale perspective and adding productivity to the value chain of e-Commerce retail.
“We want to figure out how to get the product faster to our customer and making sure we understand from AI and machine learning and other techniques and weak links in our network that are driving the loss of time. AI is helping us to get to customers faster. [Its] more resilient to first customer demand and customer expectation when they are shopping on e-commerce.
“We see that user experience online is changing and there is a lot more thought given to what is shown to customers and how they interact with the web page,” Singh told indica.
“Gap Inc has been already invested in building quantitative models for 30 plus years, and as we learn new techniques, we are just implying it to see if it adds value to our account and value chain. We have adopted certain AI techniques that have been fruitful,” Singh said.
However, he also believes that it’s not easy and explaining why said, “You have to understand what you are trying to do with [AI] and you have to understand what the impact of it would be.”
“We run a lot of experimentation on our platform to understand user behavior and user interaction and then learn from that behavior and implement it to scale,” Singh said.
Pallav Agrawal, Director of Levi Strauss sharing his views on AI agrees that it’s transforming companies like Levi as well.
“AI is helping both on the consumer-facing side and the business-facing side. What we are trying is to leverage data to make a better decision and see how we can leverage algorithms to make our consumer experience better,” Agrawal told indica.
On the consumer-facing side, he said, “The company has been trying to build applications that help our consumers shop online and they have launched a ‘recommendation system’ [which] helps customers find products which are relevant to the products [they] are looking at.
“We also use a lot of deep learning and our goal is primarily around making the consumer experience better,” said Agrawal who leads a data science team “At the end of the day it’s about using the right tool to solve the problem.”
“If you have browsed a bunch of jeans, then we recommend different kinds of T-Shirts and a lot of we have developed is focused at consumer’s need at that stage and providing the piece of information [to] make their journey more seamless more intuitive.”
“Look, AI is technology and technology is about how we use it. So, a knife can be used to chop vegetables and a knife can hurt people as well but the thing is it’s not the knife’s fault but the people using it,” said Agrawal and added that the use of AI at Levi is fairly young and they have been using data for many years instead.
He said that there are still challenges and it’s about working in a certain domain and the right technology. The other challenge is on the Natural Language Processing (NLP) side.
“We see things are working better in the deep learning/image processing side but on the NLP side things are still challenging because the problem is that for NLP, [a field of computer science, artificial intelligence, and computational linguistics concerned with the interactions between computers and human languages] we see these model that are trained on Wikipedia corps and News corps, but the language they use and we use are different and so you say NLP model is not as stable as deep learning. Image processing on the NLP side, it’s there but harder still,” he said.
Sharing an example of how data has helped his company to be customer-friendly, Agarwal said that one of the things they realized when looked at the data is that lot of the people were complaining about belt loops, so what they were doing previously were targeting NLP and machine learning on the data, they would look at high-level reviews and applications, and NLP then were able to isolate specific items that were problem areas for customers.
“We found belt loops problems for larger and big customers and we added more belt loops,” said Agarwal with a smile. It used to be really hard and had to go through thousands of reviews and machines now read in seconds.”
“Every three months we create a new product so every three months there is a new transformation for us. […]“AI is one of them and we use AI to figure out where it suits us better, so definitely in certain areas, it is transforming because it can cut down the time and get to the signal faster than our traditional methods.”
When indica asked Arijit Sengupta, CEO of Aible, about his company’s approach on AI he said that if you look at AI today, you need to be an expert data scientist to create AI and most people do not have data scientists.
“We approach businesses and ask what matters to their business, not just creating but what they are trying to achieve and what’s the benefit of getting the prediction right and what’s the cost of getting it wrong,” he said, “We ask questions upfront.”
When asked if AI is going to play a big role in coming years, Sengupta said, “It will be huge and actually most impactful technology; the problem today is people are focused on hype rather than reality. [Companies] are still learning and do not know what questions to ask AI.”
However, Madhukar Govindaraju, founder and CEO of Numly, shared how AI is used in managing strong privacy into HR and next-generation talent management.
“Many Gen Z employees are saying they are not getting enough support from their managers to grow in their company,” Govindaraju said. Employees get disengaged and even with the booming economy many companies do not talk about the employee attrition rate [which is] down significantly.
“How do you tackle the problem and complaints within the workforce. Managers fear giving real feedback even of the best employees,” he said, “I am talking to all these companies and support, we’ve created, a privacy platform, the interaction and software can solve the problem.”
“Interaction has to be private,” said Govindaraju.
Talking about the conference Sridhar Muppalla, the host of the 7th annual Big Data Conference and the NIT Warangal brand ambassador, told indica on the theme AI, “We are creating a platform for everyone, startups as well as experienced people to come and take advantage of the conference.”
“AI is the buzz word and investors are investing their dollars so we need to bring in companies where a lot of funding is happening in the AI space,” he said.
Adding further, he said though the conference is AI focused they are also promoting the startup pitch.
“We want Startups to be successful and if they are successful, they will end up hiring more people,” said Muppalla.