{{first_name | Leader}}, welcome back. Today’s updates:
Early startups get one year of Framer’s Launch Plan free, limited spots available*
OpenAI launches AgentKit
IBM is partnering with Anthropic
Google launched its Opal AI coding app
Tools, resources, and a prompt to visualize AI workflows clearly ⬇️
Extending the offer on popular demand
Framer is the design-first web builder that helps startups launch fast. Join hundreds of YC founders already building on Framer.
Who's Eligible?
Pre-seed and seed-stage startups.
Offer available to new Framer users only.
What You Get:
✅ 1 year of Framer's Launch Plan ($900 value)
✅ Fast, professional site — no dev team needed
✅ Scales with your startup — from MVP to full site
Extending the offer on popular demand
Spots are limited — apply before October 31st to claim your free year.
OpenAI launches AgentKit
OpenAI’s AgentKit is a full-stack framework for building, testing, and managing AI agents. It brings together workflow design, evaluation, and governance in one environment, pushing agents closer to enterprise deployment than experimentation.
The launch didn’t come as a surprise; earlier SDKs hinted at it.
But the scope signals something bigger. Over the past few days, early adopters have been treating AgentKit less like a developer library and more like an infrastructure layer, a way to standardize how enterprise agents are built, measured, and controlled.
Key updates & why they matter
Unified stack: Combines Agent Builder, Connector Registry, and ChatKit, allowing teams to design, integrate, and deploy agents without context-switching across tools.
Visual workflow builder: Maps triggers, logic, and roles in a no-code canvas, bridging product, engineering, and operations.
Evaluation and optimization: Adds datasets, trace grading, and tuning tools for real-time performance visibility.
Reinforcement Fine-Tuning (RFT): Introduces feedback-driven learning to refine agent behavior based on user interactions.
Governance and compliance: Versioning, audit trails, and policy controls built in for enterprise oversight.
Things to keep in mind
Some features (like RFT and external connector support) are still rolling out.
Integrations remain optimized for OpenAI’s own models.
Current benchmarks are self-reported, not independently verified.
OpenAI is expanding its role from building models to building the systems that run them. AgentKit shows how it wants to power the full life cycle of AI agents.
If I were in your seat, I’d be asking:
How dependent do we want to be on one provider’s ecosystem?
Can our current architecture plug into new agent frameworks as they evolve?
What happens to our workflows if pricing, access, or performance shifts overnight?
In other news
IBM is partnering with Anthropic to add Claude AI to IBM's software and has released an enterprise AI agent guide together.
Google launched its Opal AI coding app in 15 more countries, making no-code app creation via prompts available globally with improved speed and debugging
Productivity Tools
Mailbutler – Draft, summarise and track emails; auto-extract contacts into tasks
Clockwise – Re-arranges calendars to defend maker-time without endless back-and-forth
Beautiful AI – Instantly formats decks so every slide stays on brand and automates design
Krisp – Cancels noise and spits out searchable meeting transcripts and notes
Get featured tomorrow: How do you use AI for business/personally? Interesting stories will be shared with 100K curious readers.
Useful Resources
How to create AI prompts that eliminate bias and improve conversions
New guide to OpenAI’s Agent Builder and deployment best practices
How Microsoft is connecting cloud, productivity, and enterprise AI
Why you shouldn’t build an OpenAI agent without Zapier
When to vibe code an app and when not to
Which function in your company will feel AI’s biggest impact by 2026?
💰 Funding
Feedzai raised $75M at a $2B+ valuation to expand its AI platform for fighting financial crime.
Phaidra secured $50M+ Series B to build AI agents that optimize industrial and energy operations.
💼 Roles in AI
🐦 Did you see this?
OpenAI just published the leaderboard of the trillion-token economy.
Here it is:
1/ Duolingo – Isaac Andersen, Senior SWE
2/ OpenRouter – Alex Atallah, CEO & Co-founder
3/ Indeed – Chris Colon, Director of AI Platforms
4/ Salesforce – John Emmons, AI Leadership
5/ CodeRabbit –— #Aadit Sheth (#@aaditsh)
5:02 PM • Oct 7, 2025
Executive AI Visualization Generator
When to use this?
You need an image or slide-ready visual that explains an AI concept, workflow, or transformation story to non-technical stakeholders.
You are an art director for a global consulting firm.
Create a clean, professional illustration that visualizes the following concept:
“How AI transforms [insert function: e.g., supply-chain forecasting / patient diagnostics / marketing personalization / customer onboarding] from manual steps to automated intelligence.”
Style guide:
Minimalist, corporate aesthetic (white background, blue/gray palette)
Use icons, data-flow arrows, and labeled nodes
No stock people, robots, or circuit patterns
16:9 ratio, presentation-ready
Output caption: ‘AI-Driven [function] Workflow’Correct Input Style:
Define your focus area: “Marketing: personalized ad targeting and budget optimization,” or “Finance: real-time anomaly detection in expense reports.”
Add context: “Target audience: senior executives reviewing quarterly transformation plan.”
Paste the prompt and run. Describe current workflows and mention 3–4 business functions and key pain points
P.S. Get more such prompts in the Prompting Playbook (free for you)
Q. If AI could generate one thing for fun, what would you pick?
Results: Over 40% of voters want AI to generate movie scripts!
Stay curious, {{first_name | leaders}}
PS. If you missed yesterday’s issue, you can find it here.