HK-based AI startup aims to help businesses boost sales

October 04, 2019 16:02
Dayta AI, led by CEO Patrick Yu, has created a cloud-based video analytics platform that tracks and categorizes customers through network cameras.  Photos: HKEJ, Dayta AI

Marketing and sales activities are increasingly relying on analytical and quantitative tools nowadays among businesses around the globe, and Artificial Intelligence (AI) has emerged as one of the most hyped ways to improve product, pricing and service customization. Amid this backdrop, Dayta AI, a Hong Kong-based AI technology startup, is focusing on tailor-made AI solutions for enterprises.

The company's co-founder and CEO Patrick Tu recently sat down with the Hong Kong Economic Journal to outline his ambition to deploy AI to help enterprises transform business challenges into opportunities with ease.

Excerpts from the interview:

Q: To begin with, can you explain briefly how Dayta AI helps businesses?

A: Our firm was officially established in May 2018, focusing on AI-powered technology solutions. With our cloud-based system Cyclops, businesses, such as merchants and restaurant owners, can provide us access to the livestream video of their surveillance cameras. Our system will provide instant analysis on the video, powered by our computer vision algorithms, to deliver analytical results on geographical distribution of customers within the area, customer behavior and emotions. Our results can help the clients gain insights on the key performance indicators, say, the in-store service and customer engagement, of their stores.

Q: There are already lots of tech firms and startups focusing on this field. What would be the competitive advantage of Dayta AI against its peers?

A: From our understanding, AI technology service providers in the market normally require their client to re-install their surveillance cameras with specific equipment and standard. That may take about two weeks and the store would not be able to be business as usual during the period, which is probably going to cost the clients some money.

We don’t need that hardware installation process. As long as there is a sufficient number of surveillance cameras installed and connected with the internet, businesses can provide us access to the livestream video, and let our Cyclops system run the AI-powered analysis, which can be done within 5 minutes.

We also provide a flexible pricing package, with a monthly fee of HK$2,000 for store owners to subscribe to our services.

Take one of our clients, a Thailand-based fashion brand, as an example. After the business provided our system the access to their livestream video of two of their retail stores, we were able to start working, remotely in Hong Kong, to create a heat map using AI, to analyze the geographical distribution of shoppers inside the store.

The analysis found that shoppers' flow in one of the stores concentrates around the entrance and cashier, where there are discounted items placed for sale. With our analysis, the client was able to revamp the sales strategy and store design, and they have now provided us access to the livestream surveillance video of their 100 stores for our service.

Q: Nowadays, with tech giants soaking up all available talent, many organizations are facing AI talent shortage when trying to deploy the hyped technology. How does your firm deal with this challenge?

A: Many of our team members are university graduates and students. We acquired our first batch of clients with assistance from one of our angel investors, then we moved our office to the Hong Kong Science and Technology Park (HKSTP), and the institution brought us another batch of our key clients.

We are now one of the incubatees of the HKAI Lab, an accelerator program fully funded by the Alibaba Hong Kong Entrepreneurs Fund and SenseTime, with support from HKSTP, which enables us to expand our business reach to the Southeast Asian market.

As for the talent recruiting, we have now 17 developers in our team, about 80 percent of whom have graduated from the Hong Kong University of Science and Technology (HKUST). We have close ties with HKUST, and we can recruit talents from the job fairs hosted by the university.

Q: With their growing popularity, AI and facial recognition technology have also raised widespread concern over privacy. How does your company protect the personal data acquired with the surveillance camera video from stores?

A: Training of AI-powered algorithms requires a huge volume of data. And for Dayta AI, 70 percent of the data in the company’s system is obtained from open source databases, such as research databases from multiple universities. Our access to personal data, like the customers’ faces captured in the video, is authorized by legal means.

We also promise to our clients that our system only records the data from our AI-powered analyses, which does not include any video containing faces. In addition, during the process, our system automatically blurs the faces captured in the video and does not conduct any face recognition or matching process.

Dayta AI ensures that its business is in compliance with Hong Kong’s privacy-related laws and regulations, as well as EU’s privacy law General Data Protection Ordinance (GDPR). 

This article appeared in the Hong Kong Economic Journal on Oct 4

Translation by Ben Ng

[Chinese version 中文版]

– Contact us at [email protected]


Hong Kong Economic Journal