Get tomorrow's brief in your inbox
Today: Zapier and Lindy are dueling for the AI automation crown while Microsoft launches Copilot Tasks to handle your busywork in the cloud. Figma partnered with OpenAI to bridge design and code, and Nvidia's Jensen Huang told investors they got the "SaaSpocalypse" completely wrong.
The AI automation war just got serious. Zapier published a direct comparison piece positioning itself against Lindy, the upstart agent-first platform gaining traction with solo founders and small businesses. The core difference: Zapier brings 8,000+ native app connections, proven enterprise scale, and a full orchestration ecosystem (workflows, forms, databases, chatbots). Lindy offers experimental features like computer use, voice calls, and web scraping, but runs most integrations through third parties and just launched its enterprise tier in late 2025.
For small businesses, this boils down to a clear trade-off. If you need reliable automation across your existing tech stack (CRM, accounting, marketing tools), Zapier is the safer bet. It charges per completed task, not per workflow step, which keeps costs predictable. If you want to experiment with cutting-edge agent capabilities or build a personal AI assistant for inbox management and scheduling, Lindy is worth testing. Just know you're betting on a newer platform without Zapier's track record at scale. Both offer free plans to start.
The broader takeaway: AI agents are no longer speculative. They're handling lead qualification, customer support, project management, and more. The question isn't whether to adopt them, but which platform fits your needs today. Start small, test with non-critical workflows, and scale what works. Lindy vs. Zapier comparison
Microsoft Copilot Tasks turns AI into your background worker. Instead of running on your device, Copilot Tasks uses Microsoft's cloud-based computer and browser to handle recurring, scheduled, or one-time jobs while you do something else. You can ask it to organize subscriptions and cancel unused ones, turn emails and attachments into slide decks, surface urgent emails and draft replies, or track apartment listings weekly and set up home tours. It asks permission before meaningful actions like payments or sending messages. This is Microsoft's answer to Claude Cowork, ChatGPT Agent Mode, and Perplexity Computer. Currently in limited preview with a waitlist, but this is the direction agentic AI is heading: tools that work independently, report back when done, and free you to focus on higher-value work. Microsoft Copilot Tasks announcement
Google launched Nano Banana 2 for faster image generation. The new model (technically Gemini 3.1 Flash Image) creates more realistic images faster than the original Nano Banana, retaining the high-fidelity characteristics of Nano Banana Pro but with quicker turnaround. It supports resolutions from 512px to 4K, maintains character consistency for up to five characters, and handles detailed prompts with richer textures and sharper detail. It's now the default for image generation across the Gemini app (Fast, Thinking, Pro modes), Google Search via Lens and AI Mode, and Flow video editing. For Google AI Pro and Ultra subscribers, Nano Banana Pro remains available for specialized tasks. All images get a SynthID watermark and support C2PA Content Credentials for verification. Available in preview through the Gemini API, Vertex API, and AI Studio. Google Nano Banana 2 launch
Figma and OpenAI integrated Codex for seamless design-to-code workflows. Users can now start in Figma or Codex and move between design and implementation without context switching. Engineers can iterate visually without leaving their coding environment, and designers can work closer to real implementation without becoming full-time coders. This follows Figma's similar partnership with Anthropic for Claude Code. The integration uses Figma's MCP server to pass design files, Figma Make, or FigJam details directly into Codex. OpenAI reports over a million weekly Codex users and a million downloads of the MacOS app in its first week. If you're building digital products, this is the workflow integration to watch. Figma + OpenAI Codex integration
Trace raised $3M to solve the AI agent adoption problem in enterprises. The YC-backed startup maps complex corporate environments and processes so agents have the context they need to scale quickly. It builds a knowledge graph from tools like email, Slack, and Airtable, then lets users prompt high-level tasks ("design a new microsite" or "develop our 2027 sales plan"). Trace returns a step-by-step workflow, delegating tasks to AI agents or humans with the specific data each needs. The company positions itself as the manager that knows where to deploy OpenAI and Anthropic's "brilliant interns." Investors include Zeno Ventures, Goodwater Capital, and Formosa Capital. While enterprise agent orchestration is crowded (Anthropic and Atlassian are building competing solutions), Trace's knowledge-graph approach to context engineering could differentiate it. Trace funding announcement
Zapier outlined a four-stage AI maturity framework for businesses. Most companies start at Stage 1 (individual AI experiments with no central visibility), move to Stage 2 (connecting AI tools across systems), then Stage 3 (coordinated AI orchestration), and finally Stage 4 (AI embedded into core operations). The key insight: moving between stages doesn't require a major overhaul. Start by inventorying which teams use which AI tools, identify high-value workflows worth connecting, and introduce an integration layer (like Zapier) to reduce manual handoffs. Signs you're ready to advance: multiple teams want to connect AI tools to existing systems, you're manually moving data between tools multiple times per week, and leadership is asking what the organization is getting from AI investments. This is a useful diagnostic for understanding where your business stands and what comes next. Zapier AI maturity framework
Norway's $2 trillion sovereign wealth fund is using Claude AI to screen investments for ethical risks. NBIM, which manages the fund, deployed Anthropic's Claude model in November 2024 to provide governance and sustainability insights to portfolio managers. AI screens all companies on their first day of entering the equity portfolio, scanning public information beyond what data vendors cover. Within 24 hours, the system flags companies with potential links to forced labor, corruption, or fraud, often before the information hits international media. NBIM said it identified and sold risky investments before the broader market reacted, avoiding losses. This is particularly valuable for smaller companies in emerging markets where news is limited to local-language outlets. If a $2 trillion fund trusts AI for ESG risk monitoring, smaller firms should consider how AI can improve due diligence and compliance workflows. Norway fund uses Claude AI
Zapier offers a free plan and more affordable entry-level pricing than Lindy. Zapier charges only for completed tasks, not every workflow step, which keeps costs predictable. Lindy's individual products are powerful, but it lacks Zapier's proven track record at enterprise scale (Zapier is used by the majority of the Fortune 1000). For solo founders and small businesses testing AI automation, start with free plans from both platforms and see which fits your workflows before committing. Zapier vs. Lindy comparison
Jensen Huang said investors got the "SaaSpocalypse" wrong. Nvidia's CEO told CNBC that fears of AI making SaaS obsolete are overblown. AI agents won't displace traditional software tools like browsers and Excel; they'll "use these tools on our behalf and help us be more productive." Market watchers agree. Software firms that pivot from service-based models to outcome-based models (deploying AI agents for customers) will capture market share in the next two years. Analysts at HSBC said software is already eating AI, and the lion's share of value from AI will be generated within the software sector. For small businesses, this means your existing SaaS tools aren't going away. They're getting AI superpowers. Don't panic-switch providers; watch how your current vendors integrate AI and evaluate based on results. Jensen Huang on SaaSpocalypse
Mistral AI partnered with Accenture in a multiyear deal. The French AI lab will work with the global consulting giant to develop enterprise tech powered by Mistral's AI models for clients. Accenture will also become a Mistral customer and roll out the technology to its employees. This follows similar deals between Accenture and OpenAI/Anthropic. AI companies are increasingly partnering with consultants to drive enterprise adoption, betting that consulting firms can bridge the gap between cutting-edge AI and practical business implementation. For SMBs, this trend matters: if you work with consultants or advisors, ask them which AI platforms they're integrating and how they plan to deliver ROI. Mistral and Accenture partnership
Google absorbed Intrinsic, its AI robotics moonshot, back into the main company. Intrinsic, which launched as an independent Alphabet "Other Bet" in 2021, pitched itself as an Android-like layer for robotics. It will now operate as a distinct group within Google, working closely with Google DeepMind and utilizing Gemini AI models and cloud resources. This signals Google's strategic shift to pull experimental projects closer to its core business as it doubles down on physical AI. For businesses, the takeaway is that robotics AI is moving from experimental to practical faster than expected. If your operations involve repetitive physical tasks, start tracking how AI-powered robotics evolves over the next 12-24 months. Google absorbs Intrinsic
Burger King is launching an AI chatbot called "Patty" in employee headsets. The OpenAI-powered assistant helps with meal preparation and evaluates customer interactions for "friendliness" by recognizing phrases like "welcome to Burger King," "please," and "thank you." Managers can ask the AI how their location is performing. The system integrates with Burger King's cloud point-of-sale, alerting managers if machines are down or items are out of stock. The company is also testing AI drive-thrus in fewer than 100 restaurants but says "not every guest is ready for this." Patty will pilot in 500 restaurants by the end of 2026. This is a reminder that AI isn't just for knowledge work. Service industries are deploying AI for operational efficiency and quality control. If you run a service business, consider where AI could monitor consistency, reduce errors, or improve customer experience. Burger King AI assistant
Automate Google Business Profile review responses with AI. Use Zapier to trigger on new reviews, send the review text to ChatGPT (or AI by Zapier), and automatically post a professional reply. This keeps your response rate high without manually checking your profile daily. Set up a simple prompt that includes your business tone and key information, and the AI will draft replies consistently. You'll still want to review and tweak for complex or negative reviews, but this handles 80% of the workload. Google review automation guide
Combine your code and design examples to build new tools faster. AI coding agents work best when you give them working examples to combine. Keep a collection of snippets, tools, and prototypes you've built (or found), then prompt the agent to merge two or three into something new. For example: "Here's how to OCR an image with Tesseract.js, and here's how to turn a PDF into images with PDF.js. Build a browser-based tool that OCRs PDFs." This approach cuts development time and gives the agent clear, working context to build from. Simon Willison on hoarding code solutions
The automation platform wars are heating up. Zapier is defending its enterprise moat while Lindy carves out a niche with solo founders and AI tinkerers. Microsoft, Google, and OpenAI are all racing to ship agentic AI that works independently. The "SaaSpocalypse" panic is fading as the market realizes AI won't kill SaaS, it will transform it. Watch for more consulting partnerships (OpenAI, Anthropic, Mistral are all going this route) as AI companies try to crack enterprise adoption. For small businesses, the practical takeaway is simple: start testing AI agents with low-stakes workflows this quarter. The platforms are ready, the tools are real, and the competitive advantage goes to whoever moves first.