Summary
Google has been using AI in Search for decades. However, in recent years, advancements in AI have resulted in huge leaps forward in how Google uses AI in Search.
On May 10, 2023, Google announced that it was supercharging Search with generative AI: “With new generative AI capabilities in Search, we’re now taking more of the work out of searching, so you’ll be able to understand a topic faster, uncover new viewpoints and insights, and get things done more easily.”
To mitigate some of the risks and limitations associated with generative AI and Large Language Models (LLMs), they started with an experiment in Search Labs called Search Generative Experience (SGE). It was an opt-in experimental environment that was initially scheduled to run through the end of 2023 but, at the time of writing, is still available.
Experiment participants are presented with a “Get an AI-powered overview for this search?” prompt:
Upon clicking the generate button, an AI-powered overview is generated:
While some of the generative AI results overlap with traditional Search Engine Result Pages (SERPs), they often contain differences. For example, in the SERP above, The Law Offices of Joseph C. Bird, PLLC appears in the traditional result but not in the SGE results. If you look closely, you’ll also likely recognize several other differences, most notably:
I recently had the opportunity to discuss SGE with Tom Armitage of Smith.ai. We discussed more SGE examples in the context of law firm searches. For informational queries, the differences between traditional results and SGE results are more profound:
If you’ve been tracking changes in Google Search, these SGE results might remind you of featured snippets. In fact, many search engine optimization (SEO) professionals joked that SGE was just a glorified feature snippet. However, perhaps the most interesting changes are the follow-up prompts.
In my opinion, this is likely to have a profound impact on searcher behavior. Instead of searching once and scrolling through listings, users are more likely to click and ask follow-ups. This is likely to signal a fundamental shift in how users interact with Google altogether. In fact, this is reflected in Google’s SGE documentation:
“When appropriate, SGE will show an AI-powered snapshot to help people quickly get an overview on a topic, with factors to consider and helpful information. These snapshots serve as a jumping-off point from which people can explore a wide range of content and perspectives on the web. SGE will show links to resources that support the information in the snapshot, so people can check the information themselves and explore further. This allows people to dig deeper and discover a diverse range of content, from publishers, creators, retailers, businesses, and more, and use the information they find to advance their tasks.”
Lawyers seeking to future-proof their visibility in Google Search ought to take note of how Google is showing information for relevant search journeys and, most importantly, from where they appear to be sourcing relevant information.
But this is really just scratching the surface of where Google Search is headed.
On January 17, 2024, Google announced Circle to Search and an update to multisearch. Notably, when you point your camera and ask a question, multisearch will show results with AI-powered insights. While the examples provided may not at first seem relevant to law practice, imagine if users could point their cameras at life-legal documents (e.g., parking tickets, leases, etc) and get answers to more complex or nuanced questions. Smart law firms are already preparing for this future by finding ways to be included in these results when users dig deeper into supporting links.
Then, in February 2024, Google announced the rollout of Gemini, their most capable AI model. Despite inaccuracies in some historical image-generation depictions, Google is all-in on Gemini. At the time of writing, Gemini is available in both free and paid versions. I’m currently subscribed to the paid version. Here are some Gemini results:
Notice that this result appears to have been generated entirely from Google Maps. The listings are stripped down to business names with links to websites and reviews. Here’s a portion of our filing for divorce query in Gemini:
At the bottom of the result there is a prompt to “Double-check response.” When a user clicks on that, Google “checks” its results:
In this example, we can see that Google Search found content that’s likely similar to what Gemini generated on this law firm’s website. However, it also notes that the link provided is not necessarily what Gemini used to generate its response. While we don’t know whether Gemini relied on this page to generate the result, it’s worth noting that the site listed does use FAQ schema markup on its page. Did that impact Gemini’s understanding of the content? We don’t know.
Now that we’ve discussed SGE, Gemini, and the role of generative AI, let’s explore what I believe are the key takeaways for lawyers.
First, get familiar with relevant SGE results related to your practice. What changes do you see? What sites tend to be listed? Having tested many queries, I can tell you that major legal directories and review sites commonly appear in SGE results. If you rely on local organic search traffic for new clients, I encourage you to prioritize getting your information accurately listed on these sites.
If you haven’t done so already, it’s imperative that you find ways to deliver remarkable experiences to get reviews from clients but also from anyone who has an experience with your firm.
Second, I recommend that you research relevant follow-up prompts and make sure you have content on your site that is responsive to these prompts. Publish answers to frequently asked questions. While Google is not currently showing FAQ markup, I still believe it’s worthwhile to include appropriate structured data markup to help the machines understand information about your pages.
Think of the types of things users might use multisearch to understand. Be sure your web presence includes your firm’s perspective on these types of queries.
Third, prioritize content creation beyond the written word. As Liz Reid, vice president and general manager of search at Google stated in an interview for Wired: “In my mind this is about taking search from multi-modal input to really doing multi-modal output as well.”
Finally, rethink your firm’s attribution models. In the past, lawyers have tended to focus on last-click, direct-response attribution. Instead, lawyers should consider attribution models that take into account both quantitative and qualitative acquisition data. As user behavior on Search continues to evolve, traditional attribution models will be too limited in understanding your next clients’ journeys to finding you. Firms that are asking, “How did you find us,” are already seeing generative AI tools showing up (e.g., ChatGPT).
I foresee a future in which Search delivers generative experiences that intertwine human expertise with a variety of media. The more Google understands your reputation for knowledge, skill, experience, and excellent service, the more likely it is to surface you in its results. The more scarce its data is about you, the less likely it is to surface you as an expert to those who are looking for your expertise.
