The Gist
- More Google, more AI. Google announced a host of AI model variants at its annual developers conference, Google I/O 2024.
- What's new with Gemini? The Gemini app introduces new features that incorporate multimodal sources into its responses.
- Marketing message with search. Marketers with search strategies should pay attention to consumer adoption of the Gemini App as a signal for future search strategies.
Google at its at its annual I/O conference today unveiled a series of groundbreaking AI advancements, including the Gemini app, set to transform the future of online search by integrating AI-driven recommendation engines that promise more contextual and multimodal responses, fundamentally changing how marketers approach search strategies.
Here's a full list of the Google I/O announcements.
Google I/O 2024 Unveils AI-Driven Search Innovations with Gemini App
Much of Google I/O 2024 presentations centered on upcoming features scheduled for now through the summer, such as Gemma2, the next generation of its small language model Gemma. All of these features expanded tokenization and parametrization.
Enter search with AI. Google improvements are meant to enhance previous AI releases, such as with the Gemini App. The Gemini App helps users build context around their smartphone prompts.
Search is changing to incorporate AI as a search engine variant called “recommendation engines.” AI models are recommendation engines that craft responses to prompt queries. Those responses offer more value by incorporating modals and data types such as PDFs, JPEG image files, MP4 video files, or pasted media like lines of code.
The change will impact how marketers strategize to have customers discover their products and services online. Marketers must plan more contextual coherence for their online media. When a customer asks a query in an AI app — just as they would in a search engine — the recommendation engine will consider the best online media that it “thinks” is a reliable modal for its answer.
Related Article: Will Search Be Generative AI or Blue Links? Actually It's Both
Gemini App Utilizes Astra for Enhanced AI Responsiveness
The Gemini app operates similarly, relying on Astra, a project development framework designed as an advanced seeing and talking responsive AI agent. Astra allows the Gemini App to access the smartphone’s camera per the user’s permission so that the app can incorporate real-time camera images, identify items mentioned in the user’s AI prompt and then craft a prompt response based on the identified items.
In an I/O video, a developer walks around a desk, pointing at and picking up various items. The developer then asks Gemini a prompt mentioning the item. The app responds with options that incorporate the item in its considerations. This capability is a benefit of a multimodal model.
A key aspect that makes the multimodal model-powered search possible is increased tokenization capacity. Google announced at I/O it is increasing the multimodal window capability in Gemini. This means the text window where users craft their prompts can process up to 1 million tokens. An increase to 2 million tokens is coming later in the year. This means users can upload a multitude of files into their query and receive a very speedy response. The AI engine will treat each as a modal, explain it all in its responses, and place a context around the modals when it gives answers.
One slick hat trick Google demonstrated through the I/O was how the underlying AI engine responded to interruption. This means users no longer have to wait for a response to complete before iterating with a new prompt.
The capabilities complement two additional features emphasized — multistep reasoning and AI overviews. Each of these summarizes the modals involved so that users do not have to iterate extensively to explain to the AI model what they are looking for.
The Gemini era is here, bringing the magic of AI to the tools you use every day. Learn more about all the announcements from #GoogleIO → https://t.co/ZQaDrmItuh pic.twitter.com/9Tu1C6WNqQ
— Google (@Google) May 14, 2024
Pump the Brakes: SEO Strategy Isn't Going Anywhere
Search and SEO as we know it today will not go away next week. Keywords remain valuable. If you are launching a site, you still need to plan for the keywords that users will use in a search.
What will happen is that these Gemini changes will trigger how the industry shifts to incorporate the changes in SEO plans. The most obvious change is the role of the search bar. With a stronger emphasis on Gemini in the smartphone, users will rely more on its prompt bar than the standard search bar. If the adoption of the Gemini app grows significantly, then the prompt bar will become the search bar.
This means the value of keywords will shift to include a broader context than before. Marketers saw the first shift when search queries incorporated mobile search markers, like “restaurant near me” rather than just the word “restaurant.” The added assistance capability of the Gemini app, like the ability to identify objects in real-time through the camera, means the context around keywords has become more elaborate.
This effectively means memory — how AI recalls certain details — is an agent in a search context. The top of SERP becomes less of the main way customers notice your product or service. An image of it can suddenly be used in an AI prompt.
This also implies that keyword and SERP strategies must reimagine queries as AI-crafted layered onto human-powered search. Websites, web apps, product landing pages, and YouTube videos are still essential to search. But as the adoption of AI apps as recommendation engines rises, discovery through Gemini will rely on how beneficial they are to queries not completely centered on keyword appearance.
Related Article: The Search Engine Game Is Changing for Marketers, Content Creators
Google Sidesteps Competition with Gemini Features
Google may have dodged a significant competitive bullet with its search features. There were industry rumors that OpenAI planned to introduce a search component to its popular ChatGPT solution. But Sam Altman sent a tweet eliminating that thought just before OpenAI’s prominent announcement of ChatGPT-4o.
not gpt-5, not a search engine, but we’ve been hard at work on some new stuff we think people will love! feels like magic to me.
— Sam Altman (@sama) May 10, 2024
monday 10am PT. https://t.co/nqftf6lRL1
Google's more subtle announcement of search-related Gemini features is much more influential. Google still commands market share for search. There is still tremendous value in today’s search environment, evidenced by the revealed $20 billion annual payment from Apple. The immensity of such value makes forecasting complex (Alex Kantrowitz’s article delves into the complexities of search).
With that said about search, Gemini’s improvements have become Google’s smartest offering. Search users now have a viable alternative to their own smartphone. If they have a strong preference for using Google, then Google has revealed the most valuable AI play to date, one that retains that customer and does not lose them outright to fierce competitors like ChatGPT.
In the meantime, Google I/O 2024 has more to announce. Overall, Google’s Gemini app announcement may have said it all for the venerable tech giant.