The Center for Data Innovation spoke with Ofer Tziperman, chief executive officer of Anagog, an Israeli company that enables mobile phones to understand users’ real-world behaviors and real-time context to craft personalized customer experiences and increase retailer engagement. Tziperman discussed how this technology empowers consumers to pull personalized offerings from retailers when they have the time and attention to engage with it. At the same time, it allows retailers to anticipate consumer’s needs and provide relevant, personalized offers that are tailored to customers’ needs.
This interview has been edited.
Sujai Shivakumar: What is the primary problem Anagog is trying to solve?
Ofer Tziperman: Until now, brands and advertisers collected huge amounts of highly personal data from as many sources as they could get their hands on, so they can build a detailed and accurate user profile to help them target the right user with the right offering.
At Anagog, we decided that we should turn the traditional cloud-based concept on its head. Instead of the cloud collecting tons of private data and processing this data to push offerings to the relevant users, we want to allow the smartphone itself to learn its owner—the user—and based on each individual user’s profile and real-time context, allow the smartphone to pull the relevant offering from the cloud, but without revealing any identification to the cloud in return.
Think about the following example: You walk into your hotel restaurant for breakfast. You have two options: you can sit at a table and be served by a waiter and then you are fully identified by the waiter who needs to serve you with your order, or you can take your plate and go to the buffet. The result is exactly the same, you have on your plate what you want, but with the buffet, you did not need to identify your choices to the waiter. You just pull the food anonymously to your plate.
We ended up using edge AI to build the brain in the smartphone rather than in the cloud. And we developed the first-ever personalization engine that can use all the available data sources in the smartphone to understand both the user profile—micro-segments—and their real-time context—micro-moments—all inside the smartphone, without the need for any support from the cloud. So personal data never needs to leave the smartphone.
Shivakumar: How can data from smartphone sensors anticipate customers’ movements and needs?
Tziperman: In our smartphones, we have many different sensors, such as an accelerometer, barometer, GPS, Wi-Fi, Bluetooth, magnetometer, and more. By analyzing the different signals generated by the different sensors, and by fusing them together in the smartphone itself, the AI algorithms allows the smartphone to gain insights about its user’s activities. For example, analyzing accelerometer signals can provide insights about whether the user is walking, running, or driving. Analyzing multiple sensors together can provide insights if the user is on a bus, a train, or in a car. And since it is all done in the smartphone, all this valuable information can be understood in real-time.
In addition, there are options to understand the lifestyle of the user. If they are frequently visiting golfing ranges and clubs, they are a “golfer.” If they are frequently visiting shopping malls, they are a “shopper.” If they are running several times a week, they are a “runner,” and if they visit schools for a short time in the morning and in the afternoon, they are probably a “parent picking or dropping a child”. Furthermore, the edge AI can also investigate the host app to use any information already known to the app such as the user’s age, gender and real-time context of actions taken in the app.
For the first time ever, the on-phone edge AI is using this entire data set only in the smartphone, without sending it outside to anyone, not to Anagog and not to the app owner. This is a huge revolution.
Shivakumar: How can this information empower consumers?
Tziperman: Edge AI puts the user back in full control over his or her own data. Brands and advertisers can reach out to the users with a relevant offering without knowing who they are. The brands just need to define the rules of their campaign. What do they want to offer, to which audience, when, and where? The smartphone will then anonymously pull these campaign rules, and the edge AI, on the smartphone, will take over to decide if to show it to the user, and when and where.
Customers can finally receive greater personalized offerings, and they can finally receive them exactly at the right time, when they have the attention to engage with it. This is a big improvement from the “spray and pay” current attitude we see in the market today.
Shivakumar: How can retailers use this information to better reach customers?
Tziperman: So far, retailers have been limited with the information they collected about their customers. Limited basically to the purchase history with their brand, in store and online. If a customer has a pet but they never bought pet food, the retailer will not even know that they could offer this customer relevant pet food options. If a customer is a jogger, they would not know that they could offer this customer running shoes.
In recent times the stringent threat of, or tightening of, legal and privacy regulations such as GDPR, kept retailers away from the collection and usage of personal data. However, even those retailers who had, and still do, collected data, never really knew what to do with it. So, they collected it, and stored it, and that’s about that. They mainly absorbed the cost and the legal exposure.
Furthermore, most of the retailers don’t have any idea about the real-time context of their customers. This means they really couldn’t reach out to them when they were in the vicinity of the shop, and they couldn’t tell a customer on their way to pick up in-store what was purchased online.
By empowering the retailers’ mobile apps, retailers suddenly get to engage their customers based on better hyper-personalized data, without actually collecting such data—and all while reaching their consumers at the right time and place, in real-time. All without any legal exposure, and with minimal cost, thanks to the computing no longer in the cloud, but rather in the privacy of each individual customer’s mobile phone.
Shivakumar: Can you give examples of other applications of Anagog’s products?
Tziperman: Hyper-personalization fits every app that is seeking greater engagement with their users, including retail, banking, insurance, telecommunications, such as mobile operators, gaming, and automotive, to name a few. For example, many car manufacturers are looking to become mobility service providers, combining transportation services from public and private transportation providers through a unified gateway that creates and manages the trip. Car manufacturers can market this service only if they can provide relevant, personalized offers to customers, and this, in turn, can be done by the smartphone learning the customer’s next mobility needs.