OpenAI for Business

AI isn’t replacing design - it’s removing friction. In OpenAI’s Executive Function series, Figma’s Head of AI Products David Kossnick shares how AI is embedded across Figma, from text editing to Figma Make, turning prompts into production-ready code while keeping human judgment and creativity at the center. Read more here: https://lnkd.in/geE_k2kg

We just shared a guide on how engineering teams at OpenAI are getting the most out of Codex. It highlights how developers can use AI to code faster, stay in flow, and build with less friction: 🧠 Real-world use cases: debugging, refactoring, migrations, performance tuning ⚡️ Tactics for triaging faster, scaffolding new features, and keeping momentum up 🛠️ Best practices for prompting, environment setup, and scaling Codex across a team 📝 Sample prompts and stories directly from engineers using Codex in production Codex is still in research preview, but it’s already helping developers ship better code more efficiently. 👇 Check out the full guide here 👇

Startups building in Europe often face a tradeoff: move fast with new technologies or meet strict compliance standards. Wonderful is doing both - using OpenAI’s latest API models with EU data residency to build for enterprise at speed. With OpenAI's EU data residency: ✅ Data is stored and processed in the EU ✅ Zero data retention — nothing is logged ✅ No extra infra required More on data residency and controls in Europe: https://lnkd.in/dRBBSZW8

In the fast moving world of AI, it's a huge advantage to be able to systematically evaluate and leverage the strengths of new models as they become available. To do this robustly, you need rigorous offline evals and the scale to run meaningful A/B tests. You need to build a deep understanding of the strengths, limitations and tradeoffs each model brings to your particular application. Our modular, multi model architecture, and our ability to constantly evaluate and utilise the very best models for each specific task is one of our competitive advantages, and sees our product Fin lead the market with the best performance and AI driven customer experience. We love the work OpenAI are doing, and have shared with them some of the detail behind our approach. Link below in comments...

OpenAI changed the world when they launched ChatGPT. At Intercom we started experimenting with GPT-3.5 immediately and shipped our first features within weeks. Since then, we’ve built a sophisticated machine around how we build Fin. We shared some of what we’ve learned with the OpenAI team here.

Intercom's AI customer service agent, Fin, resolves millions of complex customer queries each month, delivering faster answers, lower costs, and consistently improving results. How did Intercom build a sustainable AI advantage? 📊 Experiment early: Hands-on testing with GPT-3.5 and GPT-4 built the model fluency they needed to move quickly. ✅ Evaluate fast: Rigorous offline evals and A/B tests let them migrate to GPT-4.1 in days, cutting costs by 20% while improving accuracy. 🧩 Design for change: A modular, model-agnostic architecture allows Intercom to evolve quickly and expand Fin into ops and product use cases. Today they're scaling beyond support into product, ops, and more. 👇 Full story in the comments below 👇

Your team moves fast. With ChatGPT, you can move even faster. In our upcoming webinar, see how small but mighty teams use ChatGPT to: 📈 Summarize research and market trends in seconds 💰 Surface cost-saving insights buried in spreadsheets 💡 Instantly find answers from tools like HubSpot and Google Drive Join us on August 13 at 10 a.m. PT for live demos and expert Q&A. Save your spot: https://lnkd.in/g_v9eS6T

We’ve extended MCP to chat in beta (alongside Deep Research) for Team, Enterprise, and Edu users globally - and Pro users outside of the EU. It lets you securely connect any internal or third-party tools such as Notion, HubSpot, and Canva so employees can access and act on company knowledge in one place. Learn how to build your own, test it securely in ChatGPT, and ask your admin to publish for your workspace when ready: https://lnkd.in/ejq9pc77

Outtake uses GPT-4.1 and OpenAI o3 to automate the detection and resolution of digital identity attacks in hours, not weeks. AI agents continuously scan webpages, app stores, ads, and social platforms, identifying threats like phishing or impersonation. They classify and act on issues automatically, while analysts get a prioritized view so they can focus on the most complex cases. The impact: ⚡ 100x faster threat resolution 📉 Zero ticket backlog 🛟 Millions saved through fraud prevention Full story here and in the comments below: https://lnkd.in/gpQKSXQm

Check out our guide to helping enterprises identify and scale high-impact AI use cases. It goes through what we’ve learned from real-world deployments to help teams move from experimentation to execution, including: 📐 A framework for mapping where AI fits across your org 🧠 Six “use case primitives” every team can apply 📊 An impact/effort matrix to prioritize what to scale 🏆 Real-world examples from Fanatics, BBVA, Match Group, and more 👇 Full guide is below 👇