- Simplifying Complexity
- Posts
- 🤯 OpenAI’s biggest reveal this year
🤯 OpenAI’s biggest reveal this year
PLUS: How to build your first AI agent workflow with OpenAI’s new Agent Builder
Alvaro Cintas
October 07, 2025
Good morning, AI enthusiast. OpenAI just took ChatGPT way beyond a chatbot, now it’s shaping up to be a full operating system.
In today’s AI newsletter:
OpenAI turns ChatGPT into an AI app platform
AMD signs massive AI infrastructure deal with OpenAI
Deloitte goes all-in on AI with Anthropic’s Claude
How to build your first AI travel agent
4 new AI tools & more
AI NEWS
At Developer Day, OpenAI unveiled new tools letting developers embed third-party apps directly into ChatGPT, from Spotify and Canva to Zillow. Users can interact with these apps entirely inside the chat window.
→ Apps SDK lets developers build adaptive, interactive, and personalized apps within ChatGPT
→ AgentKit enables drag-and-drop creation of AI agents to complete tasks on your behalf
→ Codex coding model now generally available, with Slack integration and analytics tools
→ OpenAI teases future commerce features: buy things directly from ChatGPT with instant checkout
→ OpenAI continues expanding its AI data center empire, including a 6GW AMD chip deal
Why is important: OpenAI is expanding ChatGPT beyond messaging, creating a platform for integrated apps, AI agents, and video tools, while hinting at its first consumer AI devices.
The Simplest Way to Create and Launch AI Agents and Apps
You know that AI can help you automate your work, but you just don't know how to get started.
With Lindy, you can build AI agents and apps in minutes simply by describing what you want in plain English.
From inbound lead qualification to AI-powered customer support and full-blown apps, Lindy has hundreds of agents that are ready to work for you 24/7/365.
Stop doing repetitive tasks manually. Let Lindy automate workflows, save time, and grow your business.
AI PARTNERSHIPS
AMD has signed a multi-year agreement to deploy 6 gigawatts of GPUs for OpenAI, supplementing OpenAI’s existing 10GW Nvidia deal. The partnership includes warrants for up to 160M AMD shares, vesting based on deployment milestones and stock price targets. The first 1GW rollout of MI450 GPUs is expected in H2 2026.
→ Multi-generation, multi-year GPU deployment (~2.4–3M units)
→ Warrants align OpenAI’s incentives with AMD’s market growth
→ AMD projects $100B revenue from AI infrastructure deals
Why is important: This deal validates AMD’s Helios architecture and MI-series GPUs, while giving OpenAI a second major chip partner alongside Nvidia’s 10GW agreement. It reduces single-vendor risk for OpenAI and marks a key shift in the AI hardware market, with AMD challenging Nvidia’s dominance.
AI ADOPTION
Deloitte announced a major enterprise rollout of Anthropic’s Claude AI, creating department-specific agent “personas” while continuing to build compliance products for regulated industries. This comes even as Deloitte issued a refund for a government-contracted report that contained AI-generated errors.
→ Claude agents customized for accountants, developers, and other departments
→ Aims to build compliance solutions for finance, healthcare, and public services
→ Partnership marks Anthropic’s largest enterprise deployment to date
→ Financial terms of the deal were not disclosed
→ AI integration continues despite past hallucination issues
Why is important: Deloitte’s move underscores the growing corporate reliance on AI, despite risks of inaccuracies. The rollout signals a shift toward embedding AI deeply into enterprise operations, from internal tools to regulated client services, showing that adoption continues even amid errors and challenges.
HOW TO AI
💻 How to Build Your First AI Travel Agent Workflow with OpenAI’s Agent Builder
OpenAI just launched Agent Builder, a visual tool that lets you design, test, and deploy AI workflows without writing any code. You can drag, connect, and configure nodes to create custom agents, then test their performance right inside the platform.
In this tutorial, we’ll walk through how to build your first AI Travel Agent that can plan itineraries or look up flight details, all within OpenAI’s Agent Builder.
🧰 Who is This For
Developers and non-coders building AI agents quickly
Product teams prototyping AI workflows without engineering overhead
Entrepreneurs integrating AI into apps and services
Creators designing task-specific AI assistants
STEP 1: Start a New Workflow
Begin by logging into your OpenAI Platform and selecting Agent Builder from the main dashboard. You’ll be greeted by a clean canvas where every project starts with a default Start Node.
This node defines the input variables or state variables your workflow will use. For our travel agent, the default setup is fine, we’ll keep it as is.
Once you’re ready, click the “+” icon next to the Start Node to add your first functional component.
STEP 2: Add a Classifier Agent
Now, we’ll create a classifier that decides whether the user wants flight information or an itinerary.
Click Add Node → Agent Node, rename it “Classifier”, and add this description:
“You are a helpful travel assistant that classifies messages as either ‘flight info’ or ‘itinerary’.”In Output Format, select JSON and add a property
classificationwith options:"flight info"and"itinerary".Connect the Start Node to the Classifier Node.
Add an If/Else Node with the condition:
input.parsed.classification == "flight info"If true → flight agent
If false → itinerary agent
Connect the Classifier Node to the If/Else Node to complete branching.
STEP 3: Build the Specialized Agents
Next, we’ll build the two main agents that handle the actual responses, one for flights and one for itineraries.
Start with the Flight Agent.
Click Add Node → Agent, rename it “Flight Agent,” and in the description box type:
“You are a travel assistant that helps users find real-time flight information. Always recommend a specific flight using airport codes.”
Below, enable Web Search Access so that the agent can pull up-to-date flight data from the internet.
Connect this Flight Agent node to the “If” branch of your If/Else Node.
Now, add another node for the Itinerary Agent.
Click Add Node → Agent, name it “Itinerary Agent,” and type:
“You are a travel assistant that creates concise and well-structured travel itineraries.”
Connect this node to the “Else” branch of the If/Else Node.
At this stage, your workflow should visually look like this:
Start Node → Classifier → If/Else → Flight Agent / Itinerary Agent.
You’ve now built the full logic flow for your AI Travel Agent.
STEP 4: Test and Deploy Your Agent
Now it’s time to bring your workflow to life.
Click the Run Preview button on the top-right corner of the screen. In the message field, type:
“What should I do in a day in Tokyo?”
The Classifier routes it to the Itinerary Agent, and your one-day Tokyo plan appears.
To add visuals to the Flight Agent: open Widget Studio, design your flight display, download the template, upload it back, and preview. Personalize prompts with colors or local times. Test with:
“Find me flights from SFO to Tokyo on October 7th.”
Finally, once everything works perfectly, click Publish at the top-right corner. Name your project “Travel Agent.” You can now deploy it directly using the Agent SDK or simply copy the Workflow ID and integrate it into your own product using Jacket.
And that’s it, you’ve just built and deployed your first AI Travel Agent entirely through OpenAI’s visual Agent Builder.
ESSENTIAL BITES
ElevenLabs released Agent Workflows, a visual tool for creating branching, interactive voice conversations.
Taylor Swift fans suspect AI was used in her Google scavenger hunt videos promoting her new album, though it’s unclear if any clips were actually AI-generated.
Google launched CodeMender, an AI agent that automatically finds and fixes software bugs using Gemini Deep Think.
MrBeast tweets that AI could threaten creators’ livelihoods, calling it “scary times,” amid the rise of AI video tools like OpenAI’s Sora 2 and YouTube’s AI editing features.
Adobe expects U.S. online holiday sales to reach $253.4B, up 5% from last year, with AI shopping set to jump 520%. Mobile shopping, buy-now-pay-later, and social media ads will also boost sales.
HOT AI TOOLS
⚙️ Agent Builder: Build & publish workflows with ChatKit & SDK
🧩 Agent Workflows: ElevenLabs’ new tool for routing chats smartly
THAT’S IT FOR TODAY
Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!
See you tomorrow :)
- Dr. Alvaro Cintas
