The digital advertising landscape is undergoing a massive shift as consumers increasingly turn away from traditional keyword search, favouring the conversational interfaces of AI to research products and navigate the web.
For marketers, this behavioural change presents a complex challenge: how to accurately measure ad performance and optimise campaigns across a non-traditional user journey. Meanwhile, publishers are bracing for the impact as AI search threatens to drastically reduce referral traffic.
To explore how ad tech infrastructure is evolving to meet this moment, FutureWeek sat down with Luke Fenney, SVP of Publishers and Platforms at LiveRamp, to discuss their integration with OpenAI, the mechanics of measuring AI search, and strategies for publishers navigating this new era.
LiveRamp recently partnered with OpenAI to bring measurement and advertising capabilities to ChatGPT. What gave you the confidence to move first in this space?
ChatGPT now has almost a billion users globally, and we are seeing a massive change in consumer behavior as large language models disrupt how people find information and research products. LiveRamp’s vision has always been to create a premiere ecosystem of partners. We already work with giants like Microsoft, Netflix, TikTok, and Google, so adding OpenAI to our network was a natural fit.
As consumer journeys are reshaped by conversational search and AI agents, marketers need connectivity and measurement infrastructure to power their marketing. Through our Conversions API (CAPI) Hub, we’re enabling marketers to make secure, server-to-server data connections that reduce signal loss and offer much more reliable measurement than purely browser-based tracking. More than 450 brands using LiveRamp can now seamlessly connect their conversion events to understand how ChatGPT performs as an advertising medium.
How does measuring conversions in a conversational AI interface differ from traditional search or social media?
Consumers interact with these interfaces very differently; traditional search relies on three or four keywords, whereas AI search involves conversational questions that are typically 12 to 15 words long. Currently, brands are pushing event and conversion data into OpenAI to understand who is being exposed to ads and what the downstream impact is on online or offline purchases. The user journey isn’t a traditional funnel anymore; it’s a dynamic research mechanism across different channels. By measuring the effects of their ChatGPT ad campaigns on conversions anywhere, marketers can immediately improve measurement and optimisation, unlocking more investment. Longer term, OpenAI will also be working on in-flight campaign optimisation, allowing their advertising engine to factor in a user’s high propensity for a product.
So this real-time optimisation allows brands to send signals – like an abandoned shopping cart – straight back to OpenAI. When that ad is ultimately served, is it contextually tied to the specific conversation the user is currently having, or does the ad decision operate irrespective of the immediate chat context?
That is probably a question for OpenAI regarding their broader decision-making engine. What we are doing is feeding in a lot of the brand’s data so that OpenAI can make a broader decision in that moment about the best ad to put in front of that user. There may be alignment with the specific conversation, but it may not have that alignment as well. Our role is building that critical AI infrastructure to feed in more real-time event data – such as a signal that a user abandoned their shopping cart – so the overall ad experience can be personalised to help them finish their purchase.
There is often consumer fear regarding privacy and personal data when interacting with conversational AI. What guardrails are in place?
To be super clear, the conversion API relies entirely on fully consented data that brands have received directly from users, such as through website interactions, in-store purchases, or loyalty programmes. OpenAI uses this consented brand data to personalise ad experiences. Where LiveRamp adds another layer of protection is in how that data is handled. By establishing secure, server-to-server data connections, we help govern responsible data collaboration. Historically, brands were forced to send raw personally identifiable information (PII), like hashed emails or phone numbers, to media platforms. Instead, we send the conversion information attached to our “RampID,” which protects user privacy in a much stronger way because the raw PII doesn’t have to be communicated.
AI search is significantly impacting publisher referral traffic. Do you see this as a point of tension, and how can publishers survive this shift?
It is absolutely a challenge; depending on the publisher, declines in search referral traffic could impact anywhere from 5 percent to 50 percent of their inventory. However, what we see on our platform is that publishers investing in authenticated user strategies – creating a value exchange where users actively log in, for example – are actually not seeing a decline in user engagement or ad inventory. Publishers like The Guardian, Netflix, The Telegraph, and Zoopla are prime examples. Our advice to publishers is to start replicating the strategies of platforms like Google or Facebook. Pivot away from relying on traditional search traffic to build massive audiences, and focus on capturing your fair share of people-based ad budgets by building a strong, engaged audience that keeps coming back.
Finally, from your seat running international platform partnerships, what changes day-to-day once the pending Publicis acquisition closes?
We expect to go through the regulatory process over the next few months and anticipate the deal will close towards the end of this year, around December or January. Moving forward, LiveRamp will continue to operate as an independent company within the Publicis technology division, alongside Epsilon and Publicis Sapient. For us, nothing changes day-to-day; our commitment to neutrality, our focus on innovation, and our dedication to delivering value to our clients will continue.
