Private Internet Access (PIA) spoke with Eugene Ho, CTO and Co-Founder of Dayta AI, about the value AI brings to the shopping experience, what are some of the biggest cyberthreats in the AI industry, and the importance of protecting user data.
PIA: Thank you for taking some time for us today. What motivated you to start your company?
EH: We started the company before graduation – we first positioned ourselves as a customised artificial intelligence solutions provider. We were on winning streaks and I particularly enjoyed the process of transforming users’ pain points into AI ideas, clearing all the tech hurdles and beating clients’ expectations in the end.
The game-changing moment was when we won the LVMH challenge with a tailor-made visitor analysis solution. Throughout the competition, we realized the lack of data proficiency in the industry and, more importantly, the potential for offline visitor data analysis. This opportunity became our inspiration and helped formulate our value proposition and ultimately gave us a competitive edge. Four years later, that idea has been deployed in over 80 shopping malls and retail chains, and many more overseas.
PIA: What do you enjoy most about working with AI?
EH: Witnessing how much AI can achieve what we cannot, be it scalability or accuracy.
Comprehensive data acquisition is not straightforward. While data is usually collected online, it could be fragmented. On the other hand, traditional offline data collection tools often require hefty up-front costs, additional labour resources and extra hardware installation, yet they fail to provide qualitative data for offline space.
We have been working on using AI to help make data analytics accessible – even for those without much technical knowledge. With AI and big data, even SMEs can benefit from digitally transforming their working environment and streamlining their business processes, hence achieving facilities optimisation and resource utilisation.
PIA: Talk about Dayta AI, how does it benefit retail shopping, and what is your signature product?
EH: The retail industry is starting to understand the importance and benefits of digital transformation, the data available to them and the way it helps them make data-driven decisions. On balance, the industry is adapting AI-assisted footfall analytics.
Our proprietary product, Cyclops, provides seamless data collection and business intelligence leveraging only existing infrastructure. With Cyclops, users can manage venues on a real-time basis and visualise customers’ behaviour. Our customised business intelligence reports also offer in-depth insights and data-driven strategies to enhance visitors’ experiences. Combining data and actionable insights, our solution can drive business performance and generate more revenue.
PIA: What are some of the biggest cybersecurity challenges with AI and how does your company deal with them?
EH: Being a cloud-native AI SaaS product puts us in the crosshair of blatant cyber attacks. As AI utilises much more data for computational purposes compared to traditional software, data comes in bulkier sizes which enlarges the scope of security.
With security as our top priority, we have built our cloud infrastructure with well-planned network isolation, implemented industrial-grade security best practices and adopted zero trust policies such as least privilege access rights to reduce attack surfaces.
We understand apart from infrastructure, users are also one of the most important factors in cybersecurity. Recently, even big names like Uber had experienced data leaks despite their safe infrastructure as their contractor confirmed multi-factor authentication for unauthorised access. With this in our mind, regular cybersecurity education for internal and external users is the key to uphold Cyclops’ data security.
On top of user-side measures, we also conduct regular security reviews, vulnerability tests and unauthorised access monitoring to make sure our SaaS product is competent in the market.
PIA: How does Dayta AI handle its customers’ data security?
EH: Safeguarding customer privacy is our top priority – our technology complies with GDPR (General Data Protection Regulation) and data protection standards of other jurisdictions. In contrast to solutions that employ facial recognition and record visitors’ identities, we use only body attributes and features that do not contain any customer identities or personal identifiable data. In addition, such collected data are also aggregated by time, which cannot be pinpointed to any single person. This minimises the incentive for unauthorised access or attacks to our database.
Moreover, for all user-provided data, the information at rest and during transit inside our VPC is encrypted, and access is restricted behind MFA upon users’ adoption. We regularly conduct penetration tests on all our user-facing endpoints to ensure the integrity and safety of our customers’ data.
PIA: Why do you think individuals and companies need to have a good VPN?
EH: A sound VPN eliminates risks such as man-in-the-middle attacks.
For companies, a site-to-site VPN ensures the connection to an external party is safeguarded, enabling secure exchanges of private or sensitive information between counterparties. On the other hand, an internal VPN renders protection for remote access to internal services even outside of the office or designated locations, which enables coworkers to contribute as usual under WFH policies.
While many individuals are very concerned about their privacy when it comes to third parties, whether or not they monetize or breach their data, they often neglect their online privacy. As data breaches seemingly become inevitable in modern life, with a sound VPN, users can securely shield their sensitive data and digital footprint and enjoy the internet with peace of mind.
댓글