17: OpenAI’s agent
Hey, it’s Marc.
Anthropic just dropped Claude for Financial Services. This isn’t just another enterprise tool. It’s the first industry-specific AI offering from a major AI Lab. And it’s a power move.
Claude, Anthropic’s most powerful model, now integrates directly with S&P Global Global, FactSet, PitchBook, Databricks, Snowflake and Morningstar.
While OpenAI chases low-margin consumers, Anthropic is going after verticals.
They already dominate coding (Opus 4). They’re capturing enterprise. Finance is next.
This is vertical integration at scale, and it’s working. If you’re a startup: Stop building thin AI wrappers.
Meanwhile:
OpenAI launched ChatGPT Agent that could gut Google Workspace.
Amazon launched AgentCore to power enterprise-grade AI agents.
Google paid $2.4B to poach WindSurf's best people after OpenAI's acquisition talks fell through.
Top Agent signals this week:
OpenAI launches ChatGPT Agent
Midjourney announces its Enterprise API
xAI launches Grok for government
Mistral upgrades Le Chat with deep research, multilingual reasoning, & image editing and launches Voxtral, an open source speech intelligence
Netflix uses AI to create a building collapse scene in The Eternaut
Thinking Machines Lab closes $2B seed round at $12B valuation
Meta acquires PlayAI
👉 Get your brand in front of 30,000+ decision-makers — book your ad spot now.
Top Boardroom Reads This Week
AI Talent Wars (51)
Meta Superintelligence – Leadership Compute, Talent, and Data (Semianalysis)
Rise of Agentic AI (Capgemini)
How to Build an Agent (LangChain)
Investing in energy to secure America's AI future (Anthropic)
The Agent Company: Benchmarking LLM Agents on Consequential Real World Tasks
How to Scale deAI: Infra, Incentives & the Agent Factor (Epic AI)
Claude for Financial Services (Anthropic)
Power Moves
AWS goes all in on AI agents
Amazon launches AgentCore to power enterprise-grade AI agents. AgentCore is a suite of 7 services to deploy and operate secure AI agents at scale — from memory to browser tools. Alongside, AWS committed another $100M to its Generative AI Innovation Center to help customers build agentic systems. [ANNOUNCEMENT]
It also launched Kiro, an alternative to Windsurf and Codex.
So what? Enterprises now have a cloud-native agent stack with runtime, observability, memory, identity, and more — all built for scale. Think Salesforce’s Einstein meets AWS security posture.
Claude gets connectors: Now works with Figma & Stripe
Anthropic just launched a Claude tool directory with 1-click integrations for Notion, Figma, Stripe, and more. Claude can now plug into your everyday stack — from Stripe to Canva — pulling live context like Linear sprints or Figma mockups to generate outputs inside your tools. Desktop support includes local apps like Prisma and Socket; SaaS connectors are for paid users. [BLOG]
So what? This turns Claude from a good assistant into an informed operator. For enterprises, it means less prompt engineering and more direct value from your existing systems.
Vendor Spotlight: OpenAI’s ChatGPT Agent
OpenAI just dropped ChatGPT Agent — the first real attempt at AI that doesn’t just think, it does. [ANNOUNCEMENT]
You can now say things like:
“Book travel for my SF offsite,”
“Brief me on client X using my calendar + the news,”
“Analyze 3 competitors and make a slide deck.”
And ChatGPT will:
Click around the web like a human
Run code + write summaries
Log into tools via secure takeover
Drop editable docs, spreadsheets, and slides
All using its own virtual computer. No APIs, no hacks.
Why it matters: The rise of AI started with conversational AI which was highly adopted for “insights.” With ChatGPT agent, OpenAI is making agents accessible to mainstream users through easy and quick integration with Gmail, Notion, Github, etc.
Benchmarks (these are wild)
Investment banking models: Beats junior analysts
Data science tasks: Outperforms GPT-4o and humans
Spreadsheet editing: 2x Copilot in Excel
Real-world web tasks: New state of the art
So what? OpenAI integrates its multiple capabilities from Operator to Deep Research at one platform offering plug and play tool for corporates. It has been trained with real-world utility which enables them to cut human effort and use AI as a co-worker.
What’s next?
Identify repetitive workflows — research, reporting, outreach, prep.
Test with ChatGPT Agent — start with one weekly task per team.
Assign a pilot owner — someone who’ll track ROI + build SOPs.
Measure by hours saved or tasks closed — not just quality.
Redesign your workflows around agents — not just humans with tools.
Top AI Vendors Market Map
👉 We’re tracking 1000s of AI and blockchain vendors so you can find the right vendors, partners, and bet on the right tech. Sign up for early access.
That’s all for today.
Talk soon,
Marc & Team