Google has significantly enhanced its search experience, rolling out upgrades globally on January 27, 2026. AI Overviews now leverage the more advanced Gemini 3 models, moving beyond Gemini 2.5, to provide best-in-class AI responses. Users can seamlessly transition from an AI Overview into a conversational chat with AI Mode by asking follow-up questions, a feature designed to offer quick answers or deeper discussions based on user intent. Robby Stein, VP of Product, noted that this natural flow from search to conversation makes the experience more helpful.
In enterprise AI, several key developments are unfolding. PVH Corp., the parent company of Calvin Klein and TOMMY HILFIGER, is partnering with OpenAI to integrate ChatGPT Enterprise across its global operations, from product creation to consumer engagement, aiming to speed up data-driven decisions. Meanwhile, Anthropic introduced new interactive apps within its Claude chatbot interface on January 26, 2026. These workplace tools include Slack, Canva, Figma, Box, and Clay, with Salesforce integration planned, allowing subscribers to perform tasks directly through Claude.
Security and data integrity remain critical concerns as AI adoption accelerates. Cloudbrink expanded its ZTNA platform with new Safe AI capabilities, launched on January 27, 2026, to secure AI traffic for hybrid workforces. This platform identifies AI traffic and prevents sensitive data leaks, offering unified visibility. Concurrently, Afiniti launched Outcome Orchestration, an enterprise AI solution for contact centers, which has already delivered $2.5 billion in measurable value to clients by unifying data and decisions toward specific business goals.
The broader impact of AI is also under scrutiny. DBS predicts that the global expansion of AI technology will bolster the ASEAN economy, particularly through increased demand for electronics and its role in the AI hardware ecosystem. However, experts like Brent Dykes emphasize that human expertise is crucial for realizing AI's full productivity potential, as AI tools typically complete only about 60 percent of a task. A significant threat, AI "model collapse," looms as large language models train on AI-generated data, potentially leading to inaccuracies. Gartner forecasts that by 2028, 50 percent of organizations will adopt zero-trust data governance due to unverified AI data, highlighting the need for continuous evaluation of AI model behavior.
AI is also making strides in specialized fields. Intuitive Surgical, a medical device maker, continues to lead in healthcare AI with its da Vinci surgical robot, with over 11,106 systems in use by the end of 2025. The extensive data from these robots could train AI to perform operations independently. Additionally, Jennifer Eberhardt at Stanford's SPARQ unit uses AI to analyze police body camera footage, studying bias and working with departments like Oakland and San Francisco to build trust and drive effective change.
Key Takeaways
- Google upgraded its AI Overviews search feature globally on January 27, 2026, now using Gemini 3 models and integrating with AI Mode for conversational follow-up questions.
- Cloudbrink launched Safe AI capabilities on January 27, 2026, expanding its ZTNA platform to secure AI traffic for hybrid workforces and prevent sensitive data leaks.
- PVH Corp. (Calvin Klein, TOMMY HILFIGER) is adopting ChatGPT Enterprise from OpenAI to integrate AI across its global operations, from product creation to consumer engagement.
- Anthropic introduced new interactive apps within its Claude chatbot on January 26, 2026, including Slack, Canva, Figma, Box, Clay, and upcoming Salesforce integration, for Pro, Max, Team, and Enterprise subscribers.
- Afiniti launched Outcome Orchestration, an enterprise AI solution for contact centers, which has delivered $2.5 billion in measurable value to clients and achieved 100 percent client retention in 2025.
- Intuitive Surgical's da Vinci surgical robot, with over 11,106 systems in use by the end of 2025, is a key player in healthcare AI, collecting data that could train AI for independent operations.
- DBS predicts global AI technology expansion will boost the ASEAN economy, supported by strong AI-led demand for electronics and a significant increase in data center capacity (390 centers, 1.7GW in 2024).
- Human expertise is crucial for realizing AI's full productivity potential, as AI tools typically complete only about 60 percent of tasks, requiring human judgment for refinement.
- AI "model collapse" is a growing concern where LLMs training on AI-generated data lead to inaccuracies; Gartner predicts 50 percent of organizations will adopt zero-trust data governance by 2028 due to unverified AI data.
- Jennifer Eberhardt uses AI to analyze police body camera footage, studying bias and working with police departments to build trust and implement effective change.
Google AI Overviews now uses Gemini 3 and AI Mode
Google is upgrading its AI Overviews search feature to use the newer Gemini 3 models, which are better than Gemini 2.5. AI Overviews will also connect more easily with AI Mode for conversations. This allows users to ask follow-up questions and have a more natural chat experience. Google aims to make search more helpful and conversational by integrating these features. The system chooses the right Gemini 3 model for each query, from Flash for simple tasks to Pro for complex ones.
Google Search enhances AI with Gemini 3 and conversations
Google Search launched new upgrades on January 27, 2026, to make searching easier. AI Overviews now uses the Gemini 3 model globally for better AI responses. Users can ask follow-up questions directly from an AI Overview. This allows for a smooth shift into a conversational chat with AI Mode. Google aims to provide quick answers or deeper conversations as needed, making Search more helpful by understanding context and user intent.
Google AI Overviews now link to AI Mode chats
On January 27, 2026, Google announced updates to its Search experience. Users can now ask follow-up questions directly from AI Overviews, which are AI-generated summaries. This lets them move into a conversation with AI Mode for complex topics. Google also made Gemini 3 the new default model for AI Overviews worldwide, promising best-in-class AI responses. Robby Stein, VP of Product, stated that users prefer a natural flow from search to conversation, making the experience more helpful.
Cloudbrink secures AI traffic for hybrid teams
Cloudbrink is expanding its ZTNA platform to secure AI traffic for hybrid workforces. AI systems create new security risks as data moves through many paths, making traditional security models insufficient. Cloudbrink's platform will apply the same access and connectivity controls to AI agents and services as it does to users and applications. This allows real-time policy enforcement based on user, device, agent, and AI context. The goal is to provide a single management console for better visibility and simpler security, avoiding complex, stitched-together tools.
Cloudbrink launches Safe AI to secure enterprise AI
Cloudbrink expanded its secure connectivity platform with new Safe AI capabilities on January 27, 2026. These features help businesses and managed service providers secure AI agents, browser-based AI services, and custom models. The Safe AI BrinkAgent identifies AI traffic and prevents sensitive data leaks based on company policies. Cloudbrink CEO Prakash Mana noted that rapid AI adoption creates new attack surfaces. The platform offers unified visibility for hybrid workforces, and Supervity AI CEO Siva Moduga praised its secure infrastructure. The updated Cloudbrink AI platform will be generally available next month.
PVH partners with OpenAI to transform fashion
PVH Corp., known for Calvin Klein and TOMMY HILFIGER, will adopt ChatGPT Enterprise from OpenAI. This partnership aims to use AI across PVH's global operations, from early-stage product creation to demand planning and consumer engagement. CEO Stefan Larsson stated it will help build brands, speed up data-driven decisions, and foster innovation. AI will enhance product design insights, planning, supply chain, marketing, and retail. OpenAI's strong security and privacy met PVH's rigorous standards, ensuring responsible AI use and supporting the PVH+ Plan.
Afiniti launches Outcome Orchestration for enterprise AI
Afiniti introduced Outcome Orchestration, a new category of enterprise AI designed for contact centers. This solution aims to close the gap between current narrowly focused AI products and measurable business outcomes. Outcome Orchestration uses AI to unify and guide contact center data, intelligence, and decisions across people, systems, and workflows toward specific business goals. Afiniti CEO Jerome Kapelus emphasized that AI only matters if it consistently and measurably improves outcomes. The company has already delivered $2.5 billion in measurable value to clients and achieved 100 percent client retention in 2025. In 2026, Afiniti plans to expand Outcome Orchestration beyond pairing to address more contact center decisioning needs.
Intuitive Surgical leads AI healthcare with da Vinci robots
Intuitive Surgical, a medical device maker, is crucial for AI in healthcare. The company produces the da Vinci surgical robot, with over 11,106 systems in use by the end of 2025. These robots already use AI, and the extensive data collected from surgeries could train AI to perform operations independently. This advancement could lead to world-class surgical care available anywhere globally at any time. Intuitive Surgical's stock is currently expensive, with a 66 price-to-earnings ratio as of January 27, 2026.
Human expertise vital for AI productivity gains
Brent Dykes argues that AI's promise of huge productivity gains needs human expertise to fully succeed. AI tools quickly create initial drafts, getting work to about 60 percent completion. However, refining these drafts into finished products requires human judgment and skills, like using Photoshop, to reach higher quality. A Stanford study found workers only used AI for about one-third of their tasks. The "Rework" and "Forge" stages, which involve editing and creating new content, heavily rely on human knowledge and context that AI lacks. Without human experts, AI's productivity benefits remain limited.
DBS predicts AI tech growth will boost ASEAN economy
DBS expects global AI technology expansion to support the ASEAN economy, providing a cushion against risks like US semiconductor tariffs. The region benefits from strong AI-led demand for electronics shipments and is a critical player in the AI hardware ecosystem. In 2024, ASEAN hosted 390 data centers with 1.7GW capacity, a significant increase from 2020. Malaysia leads in data center investment, while Singapore is a mature hub, and Indonesia, Thailand, Vietnam, and the Philippines are emerging markets. ASEAN also actively upgrades trade agreements, including the upcoming Digital Economy Framework Agreement in 2026, to further boost its digital economy.
Jennifer Eberhardt uses AI to study police bias
Jennifer Eberhardt uses AI to analyze police body camera footage and study bias. Her research at Stanford's SPARQ unit has shown unconscious links between Blackness and crime, influencing how people perceive others. She works with police departments like Oakland, San Francisco, and New York to develop and assess training modules. Eberhardt is also studying the impact of a January 2024 California law requiring officers to explain traffic stops. AI allows her to analyze broad, systematic patterns in police-community interactions at scale, aiming to build trust and make effective change.
AI model collapse threatens data accuracy and zero-trust
AI "model collapse" is a risk where large language models (LLMs) train on increasing amounts of AI-generated data, leading to inaccuracies. Gartner predicts that by 2028, 50 percent of organizations will adopt zero-trust data governance due to unverified AI data. This degradation can cause AI models to generate confident but incorrect information in tasks like code reviews or security triaging. Wan Fui Chan from Gartner noted that regulatory requirements for "AI-free" data will likely intensify in some regions. Experts Melissa Ruzzi and Diana Kelley confirm model collapse is a real concern, especially as organizations build internal knowledge bases with synthetic content. This highlights the need for continuous evaluation of AI model behavior to maintain security and data quality.
Anthropic adds interactive apps to Claude chatbot
On January 26, 2026, Anthropic launched new interactive apps within its Claude chatbot interface. These apps are mainly workplace tools, including Slack, Canva, Figma, Box, and Clay, with Salesforce coming soon. Users can now send Slack messages, create charts, or access cloud files directly through Claude. This feature is available to Pro, Max, Team, and Enterprise subscribers, but not free users. Anthropic advises caution when granting access to sensitive information and suggests using dedicated working folders for Claude to ensure data security.
Sources
- AI Overviews gets upgraded to Gemini 3 with a dash of AI Mode
- Just ask anything: a seamless new Search experience
- Google now lets users jump from AI Overviews into AI Mode conversations
- Cloudbrink Extends ZTNA to Secure AI Traffic for Hybrid Workforces
- Cloudbrink Targets AI Security Gaps with Safe AI Expansion
- PVH reimagines the future of fashion with OpenAI
- Afiniti Introduces Outcome Orchestration, Defining a New Standard for Enterprise AI
- Forget Tech Stocks: The Medical Device Maker That's Essential to AI Healthcare
- Why AI’s Productivity Promise Falls Apart Without Human Expertise
- DBS foresees ongoing global AI tech expansion to provide cushion to ASEAN economy
- Jennifer Eberhardt Is Analyzing Police Bias With AI
- AI & the Death of Accuracy: What It Means for Zero-Trust
- Anthropic launches interactive Claude apps, including Slack and other workplace tools
Comments
Please log in to post a comment.