Emerging TechnologyAI & Automation“Cloud computing is the answer”: Q+A with Google Cloud’s Pravin Pillai

"Cloud computing is the answer": Q+A with Google Cloud's Pravin Pillai

Cloud computing and AI are instrumental in helping marketers utilize their data, according to Google Cloud executive Pravin Pillai.

What does the store of the future look like? From the showroom at NRF 2019: Retail’s Big Show, you’d think we’re on the verge of a world filled with customer service robots and making payments with smiles. Pravin Pillai, Global Head of Retail Industry Solutions at Google Cloud, thinks about it a little more realistically.

Pillai believes that retailers must first learn to make the most of their data, both on- and offline, and that cloud computing is a way there. Read on for Pillai’s thoughts about the upcoming year, how artificial intelligence and machine learning can impact both the customer experience and bottom line, and how Google Cloud makes those technologies more accessible.

Google Cloud- Pravin Pillai

ClickZ: What is your aim with Google Cloud?

Pravin Pillai (PP): Our mission with Google Cloud is to democratize access to artificial intelligence. When we think about the products we build, we’re addressing a wide spectrum of customers. For customers that have data science and advanced resourcing, we have a cloud machine learning engine and products available for them to build. On the other end of the spectrum, we have so many APIs that are pre-built and pre-trained for customers who don’t have in-house data science capabilities. We also have an advanced solutions lab with a group of machine learning experts we make available for customers.

CZ: How can AI and machine learning improve the customer experience?

PP: Early on, one of the first uses was product recommendations. Brands collected clickstream data, which lends itself to having predictive models for what the retailer thinks you want to see. Now we’re seeing machine learning impact demand forecasting. One obvious one is using images captured in-store for stock-out predictions. We have a partner, Trax, with a store shelf digitization solution. Sensors and cameras capture what’s on a store shelf. They then use machine learning models on Google Cloud to predict inventory flow through the store.

On the customer support side, Ocado, an online grocery retailer in the U.K., introduced filters that tell them the severity of incoming message from customers. They identified incoming messages that didn’t need to be responded to — just people saying thank you, for example — which made them more efficient. They were able to respond four times as quickly.

CZ: Let’s take that last question a little further. How do AI and machine learning improve the bottom line?

PP: In the past, it was more of a push model where retailers determined what products would be built. The customer had limited choice, but the model turned on itself and people started demanding different things. Machine learning helps drive customer lifetime value. That’s on the customer experience side; machine learning also has a lot of use cases on the supply chain and logistics side. If retailers are better able to predict inventory demands and get the right products at the right time, they’re not left with products they can’t sell.

CZ: There are always new studies about how marketers don’t necessarily know how to really utilize AI or make sense of the data they already have. Where does a brand start?

PP: I think in general, when you want to tackle a machine learning problem, the more data you have, the better. But it has to be the right data. We work with customers on what problems they’re trying to solve. In many cases, the data sets they need are available; this helps identify them.

Google Cloud Q+A: Data challenges

Affinity analytics is a huge problem to solve. If you bought this, what else are you likely to buy? Looking at point-of-sale data with online sales data and clickstream data, you’re able to say, “What types of customers exist in our ecosystem? We can get much more granular about affinities to products and responses to offers. All those attributes help identify a cohort, which retailers can use to better target customers with products and offers.

CZ: It’s January so of course, we have to talk 2019 predictions. What do you see coming this year?

PP: More retail execs are acknowledging that cloud computing is the answer, as a platform to build new customer experiences, both on- and offline. Retailers have always known they have access to a lot of data, even if they hadn’t captured it in the past. How can they leverage all the data in the ecosystem to get real deep insights about their customers? Once they have all this data in one place and the ability to extract insights, they want to embrace machine learning to become more predictive around all parts of their value chain.

We’ve always captured all kinds of clickstream data, but in-store data is not where it needs to be. How are people moving in the store and looking at products on the shelf? Which items are selling and which ones aren’t? What are people’s sentiments? There’s so much instrumentation you can do at the store, which is the next frontier.


US Mobile Streaming Behavior

Whitepaper | Mobile US Mobile Streaming Behavior

Winning the Data Game: Digital Analytics Tactics for Media Groups

Whitepaper | Actionable Analysis Winning the Data Game: Digital Analytics Tactics for Media Groups

Giving a Voice to Your Brand

Whitepaper | AI & Automation Giving a Voice to Your Brand

Mobile Messaging Masters

Whitepaper | Mobile Mobile Messaging Masters