TikTok is speaking with advertisers this week from the Carlton during the Cannes Lions International Festival of Creativity, in the private garden courtyard at the heart of the newly renovated hotel, where it is holding open talks on everything from data privacy to AI. 

Ray Cao, TikTok’s managing director, global head of product strategy and operations, sat down with Ad Age for a frank conversation from a small terrace overlooking the Carlton garden. 

TikTok has spent the last few months navigating the high level of interest in its data operations, given the app’s ties to Chinese-owned ByteDance. The U.S. government has been seeking assurances that it is keeping all its data—whether consumer information or marketing data—separate and stored in the U.S. 

The viral platform has also been dealing with new privacy rules that have also been impacting rivals, including Meta, Twitter, Snap and YouTube. Apple's data crackdown has made the advertising businesses on these platforms more complicated. 

Despite all of this, TikTok has been on an upswing in popularity and has become a central part of brands' marketing strategies. This includes a new deal with Unilever, which is sponsoring a CleanTok content channel in Europe, capitalizing on the fascination with home-cleaning videos on the app. TikTok also is innovating ad formats for shopping, search and AI, of course.

Here is Cao’s take on it all. 

This interview has been slightly edited for clarity.

What is TikTok talking about with brands at Cannes?

Our mission has been always consistent: Just build the product based on our user behavior and also bring that very authentic experience to advertisers and help them connect the dots on those billions of views that we are talking about. So, CleanTok, TikTok Made Me Buy It, all that kind of stuff, is basically the fundamentals.

[Also read: #TikTokMadeMeBuyIt challenged by influencer marketing platform]

Can TikTok go from a video app to a shopping app?

I am talking a lot about how we build an e-commerce solution this year, and a lot of advertisers and brands ask, “Hey, what else are we going to do.” So, shopping in general, the fully closed loop environment, we’re creating that in the U.S.

That’s basically our merchant strategy, we get them on board and help them to really sell on the platform, fully closed loop. On the other hand, for advertisers and brands, we also offer the open-loop options for them. If they have their own website, we drive to their website.

What are ‘open loop’ and ‘closed loop’ in terms of shopping on TikTok?

Live shopping is a good example of a closed loop. So when a user is actually browsing through we say, this is live and happening right now. And then you can just click through to watch live. The closed-loop part is video shopping. We are onboarding merchants’ catalogs and helping them really connect the dots on the platform by surfacing these catalogs to our users.

[Rewind: TikTok tests in-app checkout]

Will TikTok shopping work in the U.S.?

We just need to tune the product into better shape. We’re super successful in East Asia in terms of live shopping. How can you bring that experience to the U.S. consumer, and also the European consumer, that’s the work we have to do.

And from an advertising standpoint, my job is to make our advertisers really successful in Q4 this year. So, a lot of the advertising products we put out there, we’re going to make sure that they’re using them.

What brands are doing shopping ads on TikTok, and how?

Advertisers do A/B testing with us to prove incrementality and to understand how they can make it work. A lot of big-box guys are leaning in, and a lot of specialty retailers are leaning in. We have done stuff with Walmart and also we have really good conversations with specialty retailers, like Macy’s is a very good one, as well, for us. They see that we’re in this.

[Also read: Top 5 brand TikTok trends to know about now]

Let’s talk about search ads and AI.

We have been talking about machine learning and deep learning for a decade now as an industry. Our job today is actually finetuning those tools. How do you actually empower the merchants or advertisers to capture those moments? I think through advanced automation we should be able to provide a better return on investment and better impact from ads. That always was something that we’re looking at, you know, automation in general.

You launched a search ad test this year, how will that change advertising on TikTok?

Search ads are only available via the search toggle [a clickable option in TikTok’s ad platform] today. Basically, it is an extension, to see how an advertiser fits into the search behavior on TikTok.

We have pretty much an exploratory roadmap right now for search with regards to a lot of new functionalities. We’re surfacing brands for even direct-response type ads. And this will be toward later this year, even next year, in terms of development.

How is search advertising supposed to work on TikTok?

Search is more about a different kind of user behavior. So if you think about in-feed environments, the For You feed, it is curated content based on your behavior. Search is a totally different animal. Search is basically the user coming to us with the things they want to see, right? But for now, we are still studying that, in terms of what are the behaviors we have seen from the user.

And how could TikTok search advertising differ from traditional search?

Are we really thinking about keyword bidding or are we thinking more about a keyword-less approach? These are the very early-stage thoughts we’re having.

The user putting in the search query, that’s the first generation of the search engine. It’s about keyword matching, but on the other hand some of the search engines are doing a lot of fully automated products. So, let the machine do the heavy lifting to give you the recommendations.

TikTok is at the center of a data security debate in the U.S., how are you talking to advertisers about that?

We hold ourselves to a really high bar in terms of data security. We are looking at basically creating a data divide from the U.S. from all the other countries, and also building similar solutions in the other regions like Europe. So the company is taking extra precautions on user data.

And brands are willing to accept this message?

They’re really trusting us and they’re continuously investing more. That’s a very good sign for us to understand that the brands are still with us.

It is an industry trend, everyone is talking about data security right now and also, especially around first-party data protections. I guess for us this definitely comes up, but it is more in general, brands say, “We are doing this with another platform, can you do this as well.” Or “What is your take on data protection?” So, these are the conversations we are having on a daily basis with advertisers.

Clean rooms are coming up as places for brands and platforms to use data in a safer way. How is TikTok integrating those?

We are actually working on different options right now in terms of clean rooms. For example, we build clean rooms specifically for our advertisers to help us, and to help them, measure. We are onboarding with third-party clean room partners, so Amazon is a part of it, and we have LiveRamp as another partner.

[Related: How Google expanded its data clean room tools]

Apple is making it difficult for apps to monetize through ads, with strict privacy rules, how does that affect TikTok?

It’s a disadvantage, but also an advantage, because if everybody is getting impacted, it is kind of a shrinking of the gap versus our peer set in the industry. That kind of really brings us to an even starting point with other competitors. So for us, we will say, let’s embrace the industry norm, or the new norm. It is actually inevitable, it is happening. It’s a good thing to protect user privacy.

The [Apple] SKAdNetwork 4.0 is one of the things that we’re looking into, right now, building a solution for that. And also a self-attribution network, we’re working with third parties like Adjust [a mobile measurement firm]. So, we can become a self-attributing network.