- AI Agents Made EASY Newsletter by Carolina Posma
- Posts
- Airtable AI agents are now available on the free plan
Airtable AI agents are now available on the free plan
& you can now train your FlowGent AI agents by adding learnings to past chats.
Hellooo , 👋
It’s been another fast-paced AI week: new updates, interesting shifts, and a few headlines you definitely don’t want to miss.
Here’s a quick roundup to keep you in the loop 👇
Table of Contents
And as always: share what you like about this newsletter, what you miss, and what you’d love to see as well. Just hit reply (also if you have other feedback)! I actually read and reply to every message. 💛💜
BTW - want to get personal help and guidance, and learn how to actually save time using EASY AI agents rather than spent weeks on trying to build one?
Join the AI Agents Made EASY community now, only a few spot lefts at the current price:
Now let’s get into this week’s good stuff 👇
🤖 Weekly AI Agent Insights
👀 The OpenAI v. Google war continues: OpenAI is readying an AI-powered web browser that “enables AI agent integrations” - aimed squarely at Chrome and Gemini.
This is a huge step in the “AI is the new UI” change: OpenAI's browser is designed to keep some user interactions within a ChatGPT-like native chat interface instead of clicking through to websites. Can’t wait to see what this will look like!
🧠 Salesforce’s “Agentforce” assistant, first piloted on help.salesforce.com late last year, just logged its one-millionth live customer conversation.
It now solves 84% of queries autonomously, with only 4% escalated to humans, showing that well-designed AI agents can quickly shoulder heavy support loads.
This “lessons learned” article is definitely worth a read, but TL;DR:
- Prioritize native brand voice by adding straightforward instructions like “frame responses as a Salesforce representative and avoid unrelated commentary”. Add enough examples to the prompt.
- Seed with rich content upfront: ingest a large help portal into your agent’s knowledge base to improve answer accuracy from the start.
- Analyze handoff points to improve – Regularly review conversations that escalated, identify sticking points, and refine prompts or add content to reduce future escalations
👀 Check This Out: AI Agent Tool Updates
FlowGent now allows you to add Knowledge Base items via their API (so you can connect ANY datasource as knowledge base!) and they have implemented their “smart learning” system. Take a look at your chat history and if you’d like the agent to respond different in the future, just add a “learning item” to improve!
Airtable rebrands itself as AI first (and allows for free AI across all plans including the free plan!) and goes all in on their native AI agents. AI agents can work at massive scale: the platform can deploy thousands of AI workers simultaneously across datasets to automate workflows that would take human teams months. Join their webinar about AI agents in Airtable on July 15th.
Also Slack is opening up their AI plans: Starting July 17, 2025, customers who did not purchase the Slack AI add-on will have access to conversation and thread summaries and huddle notes. To use other AI features, customers will need the new version of Business+ or the new Enterprise+ plan.
🤷♀️ What Else Happened In AI This Week?
Spotify’s “next big indie band” turns out to be … pure AI! The Velvet Sundown racked up 1 million monthly listeners before revealing that every song, lyric (and even the four “band members”) were generated with AI tools. Their debut album dropped 5 June; the follow-up Paper Sun Rebellion lands 14 July.
EU issues its final “General-Purpose AI Code of Practice”. Published on 10 July, the 54-page document lays out transparency, copyright-respect and safety steps that model providers should follow ahead of the AI Act’s first enforcement date on 2 August 2025.
Meta keeps raising the stakes for AI talent. Mark Zuckerberg is offering compensation packages of up to $100 million to lure senior researchers from rivals, bolstering the firm’s new Superintelligence Labs division.
📹 Deep Dive: How I build an EASY AI Agent Chatbot for SaaS Help Center [Sold Mine for $1500]
Last week we covered Instagram appointment setters.
This week, we’re moving to one of the easiest AI agents to build - and one that SaaS companies actually pay for.
In this video, I walk you through how I built a full AI chatbot for a SaaS help center.
It handles customer questions, matches the brand tone, creates support tickets when needed, and knows when to escalate to the human team.
This is a perfect starter project if you want to:
Offer AI agents as a service
Help software teams level up their customer experience
Build a support bot that feels helpful, smart, and on-brand
Key takeaways from the video:
Use the company’s help center docs as your agent’s knowledge base
Train the tone of voice using real examples (support Q&A, internal guidelines)
Add clear fallback instructions: if the bot doesn’t know → escalate via email, Slack, or a ticketing system
Create tools for the agent to send messages or create tickets (like Zendesk, Jira, or email)
Customize the widget with branding, preset questions, and page-specific popups
Stress test the agent with real (and tough) questions from the docs
Bonus: I also show how to hook up an HTTP module so your agent can trigger real actions in external systems (without coding).
Watch the video here:
⭐️ Peek Inside The AI Agents Made EASY Community
New content this week inside the AI Agents Made EASY community
Learn how to build an AI WhatsApp appointment setter:

Join AI Agents Made EASY now, only a few spot lefts at the current price:
Learn the fundamentals of AI Agents:

Personalized coaching during the Let’s Build Together sessions
In the Let’s Build Together sessions we built an automation together every week (recordings available from previous sessions) and you can ask all your questions to get you started!
Join AI Agents Made EASY now, only a few spot lefts at the current price:
