We had the opportunity to meet with Andy Pandharikar, CEO/Co-founder of Commerce.ai during the IT Press Tour at the Plug and Play Tech Center in Silicon Valley. Andy’s vision for his new company is providing self-driving commerce for consumer packaged goods (CPG) companies by training AI to understand every consumer product in the market and to gain insights on what makes a product successful, as well as identify market trends.
This is a joint project in conjunction with Walmart and the client list includes Chanel, Unilever, Coca Cola, SC Johnson, Rakuten, and many more. 90% of the content they monitor is unstructured feedback and 10% is structured data. Text, images, voice, videos have exploded online since 2014. 85% of consumer products fail in the first two years. SKU level data drives intelligence platform for CPG
Commerce.ai strives to facilitate innovation across business functions including the insights team, customer support, executives, Quality, R&D, and product management.
Commerce.ai is taking its base intelligent algorithm and applying it across CPG. Training on basic language and apply domain understanding. They started with the unsupervised understanding of 56,000 product categories recognizing a different language for each product category. They support nine languages across all 56,000-product categories.
Making products consumers want is the new way of marketing. Start with external data and then customers provide internal data. Commerce.ai looks at review data more than social media because social signals have a smaller intensity. They determine who are the people talking about a product, what are the language patterns, what is the age of the person commenting, and where are they located. They are able to do this with voice and text data.
Unilever is currently monitoring 40 product categories. Commerce.ai crawls all SKUs in the 40 categories and provides intelligence and insights on the trending attributes of the products. They are able to take multiple SKUs and provide side-by-side comparisons. This information is also used to identify white spaces – gaps in the market. They are able to tell whether ingredients have positive sentiment (i.e., is demand for pumpkin spice declining?) Unilever is able to understand what’s working on new product launches and apply to new things in R&D, as well as identify trending issues about which people are talking.
In the cookie category, they know consumers are looking for Keto-diet compatible cookies but they are too expensive.
In electronics, a hardware device manufacturer was coming out with software updates and product reviews started to go down. Commerce.ai was able to identify which software updates were causing a bug so the client could go back and fix it.
Any client will be able to use the data to tell what to create and what features it should have based on the market. Adoption typically starts with innovation teams looking for new ways to disrupt the business. Users tend to be product people – product management, data scientists.
Chanel wants to read reviews in many different languages – Chinese, Japanese. Chanel is typically not sold on sites with reviews but they want to know what their target customers are saying about competitive products
Clients are excited because they’ve been flying blind with surveys and they are now able to listen to customers in their real habitat. Moreover, there is plenty of opportunity with 5 billion unique products
Working in Plug and Play has given Commerce.ai access to clients. Their competition is Nielsen and the survey companies. Right now, they are staying focused on CPG versus services, apps, other industries through their technology that can be applied elsewhere if there is sufficient demand.
Sentiment drives sales and Commerce.ai is able to collect, synthesize, and provide the sentiment data while the client has the sales data.