Highlights
Top Insights
1. AI is no longer just a “co-worker” you guide. It increasingly acts like a wizard: producing sophisticated outputs from vague prompts, but through opaque processes we can’t follow. This marks a fundamental change in our relationship with AI: from collaboration to conjuring.
2. By outsourcing to “wizards,” we risk losing the chance to build the judgment and expertise needed to evaluate the wizard’s work.
3. We need to learn to decide when to use AI as a collaborator, a wizard, or not at all. This discernment will be a critical professional skill.
Source: On Working with Wizards (One Useful Thing)
Top News
1. Baidu has released ERNIE X1.1, a new reasoning model that rivals GPT-5 and Gemini 2.5 Pro in benchmarks.
2. Claude can now create and edit Excel, Word, PowerPoint, and PDF files directly.
3. Adobe has released AI agents powered by its Experience Platform Agent Orchestrator.
4. Eli Lilly has launched TuneLab, an AI/ML platform offering access to its drug discovery models and datasets.
5. An ultra-fast genome sequencing system delivers life-saving genetic diagnoses in under eight hours.
Additional Insights
1. AI Was Made for RevOps (BCG)
While predictive AI has long supported RevOps by enabling better segmentation, churn forecasting, and sales optimization, the real transformation now comes from generative AI (GenAI) and agentic AI. GenAI makes insights actionable by creating personalized sales content, coaching reps in real time, streamlining RFPs and contracts, and improving pipeline intelligence, while agentic AI takes automation further by autonomously executing tasks such as scheduling, lead management, and early customer interactions. Together, they accelerate deal cycles, boost decision quality, and unlock new revenue growth. To harness this potential, companies must prioritize clean data and governance, standardize processes, adopt unified AI platforms, focus on quick-win use cases, iterate rapidly, and invest in change management and AI talent. Early adopters that embed these technologies into RevOps will build smarter, more scalable revenue engines and gain lasting competitive advantage.
2. How Every Employee Can Become an Innovator (BCG)
While GenAI tools expand access to innovation by enabling nontechnical employees to build apps, automate workflows, and test ideas, tools alone are insufficient—companies must create systems that mirror evolution by fostering variation, selection, and amplification of ideas. To do this, organizations should give teams autonomy to run safe, small-scale experiments, establish guardrails for quick reversals, and build internal marketplaces where ideas can be shared, refined with experts, and logged in central systems. Successful ideas should be amplified through shared repositories, leaderboards, and dedicated knowledge curators, while R&D teams act as “scalers” to standardize or accelerate high-potential innovations. A culture shift is also essential: employees must see themselves not just as executors of tasks but as contributors to a living, adaptive system, with frontline staff playing a key role in spotting opportunities. By combining GenAI tools with robust infrastructure, governance, and cultural reinforcement, companies can unlock bottom-up, employee-driven innovation that is both scalable and resilient in volatile environments.
3. How AI Agents Are Transforming Consumer Goods (BCG)
While research suggests AI can enhance value creation by up to 800 basis points, only about 10% of companies have fully integrated agentic AI into their workflows, leaving most firms at risk of falling behind AI-first competitors. A leading beauty brand demonstrates the potential, using AI to cut through consumer decision paralysis with highly personalized recommendations drawn from vast product data and dermatologist insights—deepening loyalty while lowering costs. Looking ahead, agentic AI is set to become a baseline expectation, and companies that thrive will be those that dedicate funding, empower business units to deploy solutions quickly, and embed AI across the product lifecycle to accelerate innovation, improve efficiency, and strengthen competitive advantage.
4. Empowering advanced industries with agentic AI (McKinsey)
Unlike traditional automation, agentic AI perceives context, reasons through complex tasks, and acts independently across systems, enabling end-to-end process reengineering rather than incremental optimization. Early deployments in manufacturing, logistics, and automotive have already demonstrated benefits such as defect reduction, faster R&D, lower logistics costs, and higher sales conversion. McKinsey projects a $450–$650 billion revenue uplift and 30–50% cost savings potential by 2030, fueled by advances in LLMs, APIs, and GPU infrastructure. Realizing this potential requires more than technology—it demands reimagined workflows, a restructured digital workforce, scalable technical and data ecosystems, and robust governance frameworks to manage risks. The organizations that succeed will be those that shift from tactical pilots to enterprise-wide reinvention, harnessing agentic AI not just as a productivity tool but as a new growth engine and strategic differentiator.
Innovation Radar
1. AI Model Releases and Advancements
Alibaba’s Qwen team has launched Qwen3-Max-Preview, a trillion-parameter large language model with blazing fast responses, long-context support, and API availability, positioning it as a major competitor to leading global AI systems despite being in preview and not yet open source (VentureBeat). Alibaba has unveiled Qwen3-ASR-Flash, a next-gen AI transcription model trained on tens of millions of hours of speech that outperforms rivals in Chinese, English, and even music transcription, while supporting 11 languages with advanced contextual biasing and noise rejection (AI News). Alibaba introduced Qwen3-Next, an 80B-parameter hybrid model that is far more efficient than Qwen3, delivering superior performance while requiring only a fraction of the training cost (Qwen).
ByteDance has launched Seedream 4.0, a faster, more advanced AI image generation and editing tool that it claims outperforms Google DeepMind’s Gemini 2.5 Flash Image (“Nano Banana”) in prompt adherence, alignment, and aesthetics while offering lower pricing (SCMP).
Claude can now create and edit Excel, Word, PowerPoint, and PDF files directly from user instructions or uploaded data, turning complex projects into ready-to-use outputs in minutes (Anthropic).
Baidu has released ERNIE X1.1, a new reasoning model that rivals GPT-5 and Gemini 2.5 Pro in benchmarks while greatly reducing hallucinations (PR Newswire).
Stability AI has launched Stable Audio 2.5, an enterprise-grade model that slashes audio production from weeks to minutes with an 8-step generation breakthrough, offering faster, customizable, and commercially safe brand audio at scale (VentureBeat).
2. AI Tools and Features
OpenAI has introduced conversation branching in ChatGPT, a new feature that lets users fork chats into separate threads from any message, making it easier to explore multiple ideas without losing the original conversation (Tom’s Guide).
NotebookLM has introduced new features, including flashcards, quizzes, upgraded reports, a Learning Guide, OpenStax-powered academic notebooks, audio overviews, and expanded classroom integration, to transform passive reading into active, personalized learning (Google). Veo 3 and Veo 3 Fast now offer lower pricing, 1080p HD, vertical video support, and production-ready stability for developers building high-quality AI-generated videos (Google).
Adobe has released AI agents powered by its Experience Platform Agent Orchestrator, enabling businesses to orchestrate personalized customer experiences, optimize marketing campaigns, and customize multi-agent workflows across Adobe and third-party ecosystems (Adobe).
Apple Intelligence is Apple’s integrated AI platform that brings on-device writing tools, image generation, revamped Siri, Visual Intelligence, Live Translation, and third-party AI integrations like ChatGPT into iPhones, iPads, and Macs, with developer access via new Foundation Models—all designed to enhance apps while prioritizing privacy and efficiency (TechCrunch).
3. AI for Science and Medicine
Eli Lilly has launched TuneLab, a $1B AI/ML platform giving biotech companies access to its proprietary drug discovery models and datasets to accelerate early-stage medicine development (Eli Lily).
4. Other
OpenAI is partnering with Broadcom to develop and produce custom AI chips starting next year, aiming to meet soaring compute demand, reduce reliance on Nvidia, and follow the path of other tech giants investing in specialized hardware (Ars Technica).
Alterego has unveiled Silent Sense, a near-telepathic wearable that detects subtle brain-to-speech signals to enable silent communication, hands-free device control, real-time translation, and speech restoration for the impaired (Yahoo! Tech).
Sneha Goenka, MIT Technology Review’s 2025 Innovator of the Year, developed an ultra-fast genome sequencing system that delivers life-saving genetic diagnoses in under eight hours, revolutionizing critical care for patients with severe genetic diseases (MIT Tech Review).
