Google Improves AI Mode for Complex STEM Questions

What this means for education and marketing in Brazil

1. The timing and impact in Brazil

With the return to classes in 2025, students, teachers, and distance learning professionals will once again face, even more intensely, increasingly complex challenges in their academic lives.

These challenges include, for example, the need for more in-depth mathematical explanations and the continued application of critical thinking, especially in areas such as science, engineering, and technology.

Furthermore, these obstacles require not only technical knowledge, but also more effective teaching strategies capable of facilitating understanding in a constantly evolving digital landscape.

In this challenging scenario, Google's AI Mode emerges as a potential pedagogical ally. It delivers faster, clearer, and more accurate answers. This support is particularly valuable when the topic involves sophisticated or difficult-to-interpret academic questions.

On September 1st, the search engine journal Google released an important announcement about this feature. According to the publication, AI Mode has received significant improvements to handle complex questions in STEM fields.

Among the improvements are more concise, scannable responses with a structure optimized for quick reading and immediate comprehension. Furthermore, the announcement comes at a strategic time: the start of the school year in the US, which could also impact Brazil.

In this article, we'll look at what these changes mean for students, teachers, SEO, and digital marketing professionals in Brazil.

What is AI Mode — and how has it evolved?

History and fundamentals

• Em March 2025, Google launched, in an experimental phase, the AI Mode within the search platform (Search), using the model Gemini 2.0This version brought multimodal capabilities (text, image, voice), advanced reasoning, and the “query fan-out” system, which breaks down the search into several simultaneous subqueries and synthesizes an integrated response.
• At the conference Google I/O, in May 2025, it was announced the Gemini 2.5, incorporated into AI Mode and also AI Overviews, with new functions such as Deep Search (in-depth research), Search Live (live interaction with image) and Project Astra (dynamic visual search).

Key Features:

• Query fan-out: divides complex queries into multiple simultaneous searches, aggregating relevant information.
• Deep Search: Quickly generates official, cited, and organized reports — an ideal resource for academic research.
• Multimodality: allows interaction with text, image and voice, and the ability to analyze visual documents in real time.
• Conversation with context: from the initial response, the user can delve deeper into the topic with follow-up questions — creating a fluid and continuous experience.

The September Update: Focus on STEM

Update Details

According to Google, AI Mode has received an incremental update, which, while subtle, has already demonstrated significant improvements in complex STEM questions. Specifically, the model now prioritizes more objective answers, focusing on the essentials upfront before presenting more detailed or technical explanations.

Additionally, responses have been optimized for scannability, using lists, direct summaries, and clear language whenever possible. This makes the academic search experience more efficient, practical, and tailored to the real needs of students and faculty.

Google executive Robby Stein commented on his X (formerly Twitter) account:

“Very excited about this week's AI Mode model update. We're seeing big improvements for complex STEM questions… Overall responses should also be tighter, easier to scan and get to the point up front before elaborating.”

– Translating: "I'm really excited about this week's AI Mode model update. We're seeing huge improvements for complex STEM questions... Overall answers should also be more concise, easier to read, and get to the point before elaborating."

Contextualization: education and SEO

• This update was timed to coincide with the start of classes, highlighting Google's educational approach to AI Mode as a study assistant. Search Engine Journal.
• In July and August 2025, Google had released features such as Canvas (for planning study sessions), uploads of PDFs and images on the desktop, and agentive capacities (AI execution of actions), in addition to geographic expansion to more countries.
• They were also introduced embedded link carousels in answers, initially on desktop and in mobile testing — reinforcing the visibility of cited sites even within AI responses.

Relevance for teaching and the academic community in Brazil

Agility and clarity for complex questions

Questions from STEM (Science, Technology, Engineering, Mathematics) often involve calculations, logical reasoning, technical definitions, and multidimensional comparisons. A model that delivers:

• Initial summary straight to the point, followed by an ordered explanation (top down);
• Clear and visual answers, with bullets, tables or comparisons;

can transform how students explore and understand topics such as derivatives, chemical reactions, electrical circuits, or physical theories. Aligning the language with technical Portuguese and teaching applicability would be the ideal way to explore this potential.

Hybrid education support tool

Teachers and prep courses can use AI Mode as:

• Quick concept summary generator for review;
• Explanation tool for specific questions in the classroom or remotely;
• Assistance in subjects such as Physics (Newton's laws), Mathematics (integrals, vectors), Chemistry (periodic tables, reactions), etc.

The update demonstrates sensitivity to the academic cycle — a clear sign that Google is targeting value-added pedagogical presentations.

Implications for SEO and digital marketing in Brazil

Visibility in AI Mode — a new level of competition

As AI Mode becomes increasingly sophisticated, a new priority emerges for brands and content creators: being cited directly in AI responses. This is because, according to recent analysis, traditional visibility, based solely on the first organic links, may no longer be sufficient to ensure relevant reach.

Especially in complex searches, the model tends to synthesize information and present a single response block, without necessarily displaying all links. In this scenario, appearing as a structured and reliable source within the response becomes much more valuable than simply competing for organic rankings.

Therefore, the new strategy requires well-organized content, with clear data, objective language, and a scannable structure, ready to be interpreted and cited by AI.

Every click matters more

Google states that:

• Organic traffic volume remains stable or increasing – but clicks become more qualitative: browsers spend more time on websites, without clicking and returning.
• The rise of complex queries and the incorporation of links within AI responses increases opportunities for qualified websites to be explored — even if they are not the traditional first results. blog.google.

Recommended strategies (for Brazil):

1. In-depth and technical content, with a clear structure (titles, bullets, summaries).
2. Use of data and real examples, preferably contextualized for Brazil — and with quotes that the AI ​​can use.
3. Schedule Markup (FAQ, HowTo, Article, etc.) to make it easier for AI to understand.
4. Academic partnerships and quality curation – produce reference content.
5. Search Console monitoring and future dashboard for AI Mode — adapt quickly.

Future perspectives: where are we heading?

Continuous evolution

• Google will continue to iterate on AI Mode based on user feedback and internal metrics.
• Advanced features such as Deep Search should expand, promoting complete and automatically cited reports.

Paradigm shift in search

• The transition from link-based results to conversational answers reshapes the search experience—and redefines how SEO and marketing should conduct themselves.
• The emergence of conversational assistants (possibly even on wearable devices) could lead to searches becoming less visual and more auditory/interactive.

Google's global vision

Executives like Sundar Pichai (CEO of Google) and Liz Reid (Head of Search) have already publicly demonstrated a clear vision for the future of online search. According to them, AI Mode will be the core of increasingly intelligent, personalized, and deeply contextual search.

Instead of simply displaying links, the new search will be guided by artificial intelligence, offering comprehensive, direct, and proactive solutions aligned with the user's context. Therefore, Google aims to create a more intuitive digital environment, where AI anticipates needs and delivers more useful and functional answers.

This movement, therefore, redefines how users interact with information, and how brands and creators should position themselves in this new ecosystem.

Recommendations for brands and content producers in Brazil

Conclusion

• The current update to AI Mode, focusing on STEM, represents a significant advance for academic research. Furthermore, it clearly reflects Google's new focus: education.
• In Brazil, both educators and students have much to gain from this tool, as long as they know how to use it well and understand how it works. Therefore, training and access become key.
• At the same time, a new imperative emerges for digital marketing and SEO professionals. More than ever, it's essential to become a citable source for generative AIs.
• This requires, first and foremost, the production of quality content—that is, scannable, well-structured material with authority on the subject.
• Given this new scenario, the path forward is clear. It's necessary to invest in reliable content, adapted to technical Portuguese, and relevant to the Brazilian context.
• Meanwhile, we continue to closely monitor the evolution of this inescapable transformation in the way we search for, access, and distribute information via AI.

Richard Alquati

Growth Marketing Specialist and CMO as a Service. With over 20 years of experience in digital marketing, I began my journey creating websites in 1999 and have worked with Google Ads campaigns since 2006. Today, I work as a CMO as a Service, helping companies grow rapidly and sustainably with data-driven strategies, automation, SEO, PPC, and integrations between CRM, ERP, and customized dashboards. I have managed over R$23 million in digital campaigns, always focusing on ROI, innovation, and scalability. I primarily serve the B2B market, supporting businesses in complex and lengthy decision-making processes with customized solutions for influencers and decision-makers. I believe that growth requires strategy, technology, and precise execution—and that's what I deliver in every project.