Recently, our Search Engine Marketing team had the opportunity to attend the Google Keynote I/O and the Google Marketing Live Keynote. Not surprisingly a major trend presented, and on everyone’s mind was the integration of generative AI and search.
Here is what you need to know from these keynotes and early tests of both Google’s generative AI and Bing Chat products.
What is the current state of AI in search?
Google and Bing are currently open beta testing their own AI with chat features in their search results. Both are testing this functionality in an opt-in experimental mode before a major rollout. Bing is testing ChatGPT in Bing, called “Bing Chat,” while Google is testing their AI Bard in Google Search, called “Generative AI.”
For example, when you conduct a search with Google, you will see an AI-powered snapshot of important information to consider relative to your search query, along with links allowing for deeper exploration. A major change is that traditional search results will be shown after the AI snapshot. This is sure to cause concern among advertisers who are paying for premium placement.
Additionally, the AI will read and summarize web pages for you without the need to visit them yourself and will cite sources if you want to read the website where the information came from.
AI-integrated Search is coming fast.
Google and Microsoft (Bing) are both investing heavily in AI and are planning to add more features later this year.
If successful, AI integration will create a better user experience than a search engine alone by providing more detailed answers faster and in a conversational manner instead of an endless presentation of links. This enhanced search could achieve widespread adoption and replace a significant amount of existing search traffic. This means Google and Microsoft must work hard and fast toward owning this new search state to avoid losing future search share.
AI will potentially replace the research part of search.
As AI conversations are predicted to replace a large part of the search journey, a prospective customer will do fewer searches and will have more AI interactions before making a purchase. The implication of this will be reduced opportunities to show paid search ads before the final purchase decision.
A reduced number of total searches, but more low-funnel searches, means that search engine marketing will have a big role to play in capturing fewer, but more qualified, leads. A great example is the car-buying experience. Traditionally, a potential buyer would move up and down the funnel searching for brand and model names, ratings, features, reviews and then finally for local dealerships that may have the car they want on the lot. With AI search, that same searcher could do nearly all research on brands, models, reviews, and features entirely in the search engine and skip the car blogs and branded websites altogether. With this scenario, they can start with an AI conversation and then do the final searches for the dealership, knowing exactly what they want when they are ready. AI can’t sell the vehicle, but it can consolidate vast amounts of information and educate buyers better and faster than multiple pieces of content scattered across numerous automotive websites.
If SEO does become less effective in the coming years, then search ads will need to be more focused on clearly presenting a brand as the best option and calling for action.
AI search will need to incorporate paid search, which could cause complications.
Since development of large language models and AI is expensive, search engines will need to monetize AI-based search. So far, neither Google nor Bing have provided an explanation of how this will work. Currently, they are both running ads alongside their tests and have stated they will be closely watching the performance of ad placements throughout this experimental phase.
As shown above, Google’s current experiment shows generative AI information at the top of the fold with the paid search ads below, which is a historic change. Search ads have always been sold with the expectation of being placed at top of page/above the fold. If paid search ads are now going to lose page position prominence, it could have massive implications for the future of search. Google could try to drive up cost per click as total search volume drops due to AI, but advertisers may be very resistant to paying for ads that no longer show at the top of the page.
Another major point of contention is what AI will recommend in its search results bar. Will it be only organic results, or will it “take a bribe” and recommend sponsored answers as well? Paid AI recommendations could irreparably damage trust in search results if they are not clearly labeled as paid. However, if the placement is purely “natural” with no financial influence affecting its placement, then it will not be generating revenue for the host domain. Google and Bing will both have to walk a fine line of making sure ads are appropriately labeled but still look appealing as options alongside organic, AI-produced results.
Search Engine Optimization may eventually become AI Optimization.
Currently AI recommendations are based on the same site signals as search engine optimizations (meta-tags, backlinking, keywords, etc.). Eventually, this could split and “AI Optimization” for websites could become a new industry niche. Savvy SEO experts may begin to develop a playbook to get AI models to latch onto their clients’ websites and brands to make recommendations for them in an AI Optimization search environment. For example, if you ask an AI chat for a restaurant recommendation, you will receive a shorter list than a traditional search; so how do you get on the AI’s preferred list? Currently, AI will make recommendations based on existing website SEO, web traffic and reviews. Most AI models will recommend only two or three choices by default (unless you ask for a top 10, for example) with a short summary statement of each result. Google search results, on the other hand, pull a near infinite number of answers, and Google Maps will drop every possible pin it can.
AI search has some big risks.
There are a number of recognized risks to using AI. One risk of particular concern to search is misinformation. There could be significant brand damage if an AI processes incorrect or inaccurate information and decides to explain a product or service based on this flawed information. For example, an AI might start telling searchers that a restaurant has a menu item that was discontinued or that they are unable to meet dietary requirements they could in fact accommodate. Finding a bad review is relatively easy to do, but finding incorrect information about your business within an AI database will be virtually impossible to address and could cause massive headaches in the future.
Showing up in search could be harder for smaller or new brands.
AI will likely make it harder for smaller and new businesses to get search traction in the future. Getting search visibility for a website as well as enough attention to get the recommendation of an AI will be nearly impossible for very small, niche or new brands. Certain AI models do not consider newer websites and make all recommendations based on older data, thereby completely ignoring a brand-new company or product. For example, a new restaurant with no online presence except a Facebook page and fewer than a dozen reviews could be ignored entirely by some AI models, as opposed to a traditional search that would have at least dropped a pin on a map in the search results if the restaurant had a Google My Business page.
What should brands do to prepare for AI search?
AI-enabled search is coming. While there are many challenges ahead, there are also opportunities. As a first step, make sure you are working with search experts who are forward-thinking and not afraid of change. The next few years will bring big surprises, and doing the same thing and hoping for the best won’t work. This will be like the “mobile first” revolution that happened a few years ago. Suddenly, any website that didn’t support mobile was behind. A proactive strategy for AI-enabled search is a must.
Understand your customer journeys. Identify where there could be opportunities for AI to take over information search and gathering at different journey stages, and determine the role that search can continue to play along the path to purchase.
Make sure you have a suitable budget and a search focus on brand authority and identity. AI will only show two or three results for a query, not dozens. This will greatly increase pressure on brands to set themselves apart and be authoritative enough to catch the AI’s attention.
Finally, ensure your product or service is clearly and easily explainable on your website. AI needs to be able to summarize and explain why your product or service is valuable and a good solution to a searcher’s needs. If that data can be found and accessed easily and accurately, it could be a significant lead and sales driver in an AI-enabled search environment.
