AI News Today: Everything That Matters for Small Businesses and Creators — June 27, 2026
A lot happened in AI this week. Figma just wrapped its biggest product conference in years. Apple's iOS 27 AI features are rolling out. OpenAI retired GPT-4.5. And two major AI events are coming up in the next few weeks that are worth knowing about. Here's what actually matters and what it means for you.
What's in This Roundup
- Figma Config 2026: Motion, code layers, and AI tools on the canvas
- ChatGPT model update: GPT-4.5 retired, new model picker
- Apple iOS 27: AI features built into everyday tasks
- AI agent standard: Google, Microsoft, and others release open specification
- Runway Agent 2.0 for marketing
- Upcoming AI events: AI Engineer World's Fair and Ai4 2026
1. Figma Config 2026: The Design Tool Just Got a Lot More Powerful
What happened: Figma held its annual Config conference June 23 to 25 in San Francisco. This was one of the most significant product updates in the company's history. The theme was bringing design, code, and motion into a single workspace powered by AI.
Key announcements:
- Figma Motion: You can now build animations directly inside Figma using a timeline, keyframes, and preset animations. Before this, designers had to export work to After Effects or other tools. That step is now gone for most common animation needs.
- Code Layers: Designers can now bring actual code into the canvas as layers, test interactions, and compare design directions side by side. Teams can sync with GitHub directly from the file. This is a meaningful step toward closing the gap between design and development.
- Shaders: You can describe a visual effect in plain text, and the Figma AI agent builds a WebGPU-powered shader for you. Effects like frosted glass, halftone patterns, and particle animations become adjustable on-canvas controls rather than static exports.
- Generative Plugins: Instead of writing JavaScript, you can now describe what you want a plugin to do and the Figma agent writes it. These plugins work like native Figma tools. This removes a significant technical barrier for smaller teams.
- Figma Weave Tools: Figma's AI image workflow tool (originally called Weavy) is now accessible directly inside the design canvas. Tasks like style transfer, image generation, and asset editing are available without switching tabs.
- Upgraded AI Agent: The Figma agent now supports skills, which are reusable instructions you create once and apply repeatedly. It also connects to outside tools including Notion, Slack, GitHub, and Excel.
What this means for you: If you use Figma for any part of your creative workflow, these updates reduce how many tools you need in your stack. Animating, prototyping, and handing off to developers all now happen in one place. For freelancers and small teams, the removal of the design-to-animation handoff alone saves real time.
Some features like Code Layers are still coming soon with waitlists open. The AI tools require a paid plan and will consume Figma AI credits at general availability. Worth checking the current pricing page before building new workflows around these features.
Affiliate CTA: Check current pricing and features on the official website.
2. ChatGPT Model Update: GPT-4.5 Retired, New Model Picker Rolls Out
What happened: As of June 26, 2026, GPT-4.5 is no longer available in ChatGPT, including for custom GPTs. Existing conversations that used GPT-4.5 will continue using GPT-5.5. This retirement was announced on May 28 and applies to ChatGPT only — the API is not affected.
Alongside this, OpenAI rolled out an updated model picker for ChatGPT Business on web, iOS, and Android. The new interface simplifies how you choose between speed and reasoning depth for different tasks. Some options were renamed and Thinking Light was removed, though the underlying models and usage limits remain the same.
What this means for you: If you rely on ChatGPT for content writing, customer support drafts, or any business task, your workflow will now default to GPT-5.5 where GPT-4.5 was selected. For most users this is an upgrade in capability, but it is worth testing your most important prompts to make sure outputs still match your expectations.
The updated model picker is a quality-of-life improvement. Instead of navigating multiple model names, you now choose based on what the task needs: faster response or deeper reasoning.
Affiliate CTA: Check current pricing and features on the official website.
3. Apple iOS 27: AI Gets Woven Into Everyday Tasks
What happened: Apple is rolling out iOS 27, and the AI story this time is less about Siri and more about practical automation built into the apps you already use. The update includes AI-assisted restaurant bill splitting, automatic password replacement after security breaches, intelligent message suggestions, calendar creation from natural language, smarter Shortcuts automation, and AI-organized Safari tabs.
Most of these features run on-device, which Apple is emphasizing for privacy reasons. The experience is designed to feel invisible. AI quietly handles the task within the existing app rather than asking you to switch to a chat interface.
What this means for you: For small business owners and freelancers who work primarily on iPhones and iPads, these updates reduce the number of manual steps in daily routines. Scheduling, messaging, and password management get faster without needing any additional app or subscription.
For brands and marketers, the shift is also worth noting. Apple's approach signals a broader direction in consumer AI: users increasingly expect AI to work quietly inside tools they already trust, not as a separate chatbot to visit. Content and customer experiences designed around that expectation are likely to land better.
4. New Open Standard for AI Agents: Google, Microsoft, and Others Collaborate
What happened: Google, Microsoft, GitHub, Hugging Face, NVIDIA, Salesforce, Snowflake, and several other companies released a joint draft specification called Agentic Resource Discovery (ARD). The goal is to let AI agents locate, verify, and connect with tools, APIs, and other agents at runtime, without needing everything preconfigured.
The practical idea is that organizations publish machine-readable catalogs on their own domains, and AI agents can discover available capabilities from those catalogs as needed.
What this means for you: This is early-stage infrastructure work, and most small businesses will not interact with ARD directly. But it signals where things are heading. The AI tools you use today are increasingly going to work together automatically, finding each other and sharing information without manual setup. For business owners evaluating AI platforms, this is a reason to favor tools that support open standards over those that lock you into a proprietary ecosystem.
5. Runway Agent 2.0: AI for Marketing Campaigns
What happened: Runway launched Agent 2.0, a new tool built into its existing platform. Users can go from a simple text prompt to a full marketing brief, campaign assets, and creative variations. The system can also analyze past performance data to improve outputs and scale creative across different formats and platforms.
What this means for you: Runway has been known primarily as an AI video tool. Agent 2.0 positions it more as an end-to-end creative production platform. If you produce social media content, ads, or video marketing, this update is worth evaluating. The ability to connect creative output to performance data in one tool reduces the friction of the typical content-test-adjust cycle.
That said, Runway sits at a higher price point than many small business tools. The value depends heavily on how much visual content you produce and whether the performance analytics are actionable for your specific channels.
Affiliate CTA: Check current pricing and features on the official website.
6. Coming Up: Two AI Events Worth Knowing About
If you follow AI tool development or want to understand where things are heading, two upcoming events are worth adding to your calendar.
AI Engineer World's Fair — June 29 to July 2, San Francisco
This is the largest technical AI conference in the world by scope and sessions. The event runs across 10 parallel tracks with over 400 sessions and an audience of 6,000-plus AI engineers and senior technology leaders. Every major AI company and infrastructure provider will have a presence. If you follow AI tools closely, this is the event where announcements, product previews, and framework releases tend to come out. For most small business owners, it is more useful to follow coverage after the event than to attend in person, but the timing means you can expect a wave of AI tool news in early July.
Ai4 2026 — August 4 to 6, Las Vegas (The Venetian)
Ai4 describes itself as the world's largest gathering of business AI leaders. The 2026 edition is expected to draw over 12,000 attendees from more than 90 countries, with industry-specific tracks across financial services, healthcare, retail, manufacturing, and government. Unlike more developer-focused events, Ai4 is built around business application and ROI. There are dedicated tracks on AI agents, beginner-friendly AI sessions, a startup stage with 50 or more talks, and a dedicated AI Policy Summit.
Ticket pricing runs from around $1,395 to $3,195 depending on access level. Early arrivals on August 2 can join peer-to-peer discussion groups before the main event opens. For small business owners who want to understand enterprise AI adoption trends and meet vendors, this is one of the most practical events of the year.
| Event | Dates | Location | Best For |
|---|---|---|---|
| AI Engineer World's Fair | June 29 to July 2, 2026 | San Francisco, CA | Technical AI developers, CTOs, engineers |
| Ai4 2026 | August 4 to 6, 2026 | Las Vegas, NV | Business leaders, SMB owners, operators |
Quick Scan: More AI Updates This Week
- Google Finance exits beta with a new Android app. Users can now track portfolios, set up AI-generated pre-market briefings, and get real-time market data in one place. An iOS app is planned for later in 2026.
- Samsung brings ChatGPT and Codex to employees through a partnership with OpenAI, giving its workforce access to AI coding and productivity tools at scale.
- Meta subscription plans go global on Instagram, Facebook, and WhatsApp. Meta One plans for AI users start at $7.99 per month and go up to $19.99 per month for higher compute queries and deeper reasoning mode.
- Qualcomm in early talks to acquire Tenstorrent for an estimated $8 to $10 billion. Tenstorrent makes AI chips on the open RISC-V standard. The deal, if confirmed, would give Qualcomm a meaningful position in AI hardware alongside Nvidia and AMD.
The Takeaway for Small Businesses and Creators
This week reinforced a pattern that has been building for months. AI is moving out of standalone apps and into the tools people already use every day, whether that is Figma for designers, iOS for iPhone users, or ChatGPT for business teams.
The practical shift for small businesses is this: you do not need to adopt a new AI tool for every task. The better question is which of your existing tools now have AI built in, and whether you are actually using those features. Most people are not.
Figma's Config announcements this week are a clear example. Many small business designers and freelancers already pay for Figma. The motion tools, generative plugins, and AI agent features are now available inside that subscription. Using them is a faster path to value than buying another tool.
The same logic applies to ChatGPT, Apple devices, and the productivity platforms you already run. In 2026, the question is no longer whether your tools have AI. It is whether you have spent 30 minutes learning what it can do inside the tools you already paid for.
FAQ
Is Figma Motion free to use?
Figma Motion is available on paid Figma plans. Using the Figma agent to generate animations will consume AI credits at general availability. The feature is currently in beta. Check the Figma pricing page for your plan's access level.
Will GPT-4.5 retirement affect my ChatGPT custom GPTs?
Yes. Custom GPTs that were set to use GPT-4.5 will now run on GPT-5.5. Most users will find the change is an improvement, but it is worth testing your custom GPTs if you rely on specific output behavior.
Is the AI Engineer World's Fair open to non-engineers?
The event is primarily technical in focus, with talks aimed at AI engineers and developers. Business leaders attend, but it is not structured around business application the way Ai4 is. If you are not in a technical role, Ai4 in August is likely the more useful event.
What is ARD and do I need to know about it?
Agentic Resource Discovery is a new open specification for AI agents to find and connect with tools automatically. It is infrastructure-level work, not something most small businesses will interact with directly. The practical implication is that AI tools will get better at working together without manual setup over the next one to two years.
Suggested Internal Links
- Best AI Image Generators for Small Business and Creators in 2026
- Best AI Video Generators for Small Business and Creators in 2026
- Best AI Social Media Management Tools for Small Business in 2026
- Biggest AI Tool Updates This Week (June 2026)