A New Era in Paid Search: AI Power

After almost 25 years, Google’s paid search dominance is beginning to adapt to the disruptive new technology of generative artificial intelligence, such as ChatGPT. We don’t yet know when or how Google (as well as its smaller competitor Bing) will incorporate AI-powered interactivity into its search results, but it’s guaranteed that we will be approaching a sea change in how we discover new content on the web.

From Search Results to Conversations

A realistic scenario in the future could be that upon submitting a typical search query, rather than seeing a list of ranked search results (both paid and organic), a user would instead interact with a totally different user interface involving a human-esque conversation with a subject-specific chat box. A hypothetical “search conversation” could go something like this between a user we’ll call Mike and the search engine:

Mike: “I’m interested in new running shoes for an upcoming 5K.”

Search engine: “Hi Mike. Your dedication to a healthy lifestyle is notable. When is your 5K?”

Mike: “I’ve been putting off registering for the event, so it’s kind of last-minute. The event is in San Diego 2 weeks from now.”

Search engine: “I see. Given that information, you may consider some shoe models that don’t require an extensive break-in period. I’ve looked up the event details and there’s a 73% chance of light drizzle during the first 2 hours of the race. You may also consider shoe models that also have adaptive treads for such conditions. What is your shoe size and color preferences?”

The logical end point of the above interaction would result in the search engine recommending a very specific list of options based on the user’s criteria. The big question is in what context the results will be provided: Will they be listed during the conversation or at the end? How many options will be provided—just one or several? Will said option(s) be a combination of organic and paid results and will one have prevalence over the other? Will elements such as reviews, pictures and multimedia be included in the results?

It’s also realistic to envision that the search engine may already have rudimentary information about Mike, such as his income bracket—further narrowing shoe options on his behalf. The point of this hypothetical exercise is that as the search engine understands a user’s preferences from such an interaction, there could potentially be limited opportunities for brands to get in front of an audience unless they meet a very specific set of criteria. This means fewer searches and fewer ads.

AI-powered search would reduce the need for multiple queries as in traditional search. In this new ad environment, advertisers could be subjected to higher costs for paid listings and fewer placements as competition for visibility intensifies. With users clicking on fewer links, advertisers may consider putting additional focus on upper-funnel ads to increase brand awareness as a hedge for diminishing lower-funnel opportunities. Advertisers will also need to further refine their targeting and persona modeling to capitalize on more focused user intent.

Finally, paid search may evolve from a keyword-centric model to a persona/conversion model to encompass the more conversational nature between the user and the search engine. Auction bids would incorporate not just the value of a keyword, but the value of an interaction and its conversion potential. In other words, we may shift from not only bidding on keywords related to running shoes, but also bidding on the filtering process resulting from human-to-machine conversation.