The Best New AI Apps of September 2026 You Should Try Right Now

September 2026 has been one of the busiest months yet for artificial intelligence apps.

The biggest technology companies are no longer competing only to build smarter chatbots. They are increasingly developing AI agents that can use apps, complete tasks, work in the background, write software, analyze information, and help users manage entire workflows.

In just the last few weeks, Microsoft redesigned Copilot around Home, Code and Autopilot, Meta expanded Muse into business workflows, and OpenAI introduced Dots as a new generation of always-on AI agents.

At the same time, a growing number of smaller AI products have appeared across productivity, automation, development, content creation and business.

So, which new AI apps are actually worth watching?

Here are some of the most interesting AI apps and tools making waves in September 2026.

1. Meta Muse

Meta Muse is arguably one of the biggest new AI apps of September.

Unlike a traditional chatbot, Muse is designed to act as a personal AI agent that can perform tasks on behalf of the user.

It can browse websites, interact with applications, work with connected services and continue working on longer tasks.

Meta recently expanded Muse with new integrations specifically aimed at small businesses.

These integrations include platforms such as:

Shopify
Canva
QuickBooks
Notion
Slack
Stripe
Dropbox
Asana
Figma
Zoom
Klaviyo

This allows Muse to work across multiple business applications instead of operating as an isolated chatbot.

For example, a small business owner could potentially use Muse to help coordinate marketing, customer communication, e-commerce activities and administrative work.

Meta is also bringing Muse to its AI glasses, making the assistant increasingly connected to the user's physical environment.

Why Muse Matters

Muse represents the shift from:

"Ask AI a question."

to:

"Give AI a goal."

That difference could define the next generation of AI applications.

2. OpenAI Dots

OpenAI entered the AI-agent race even more aggressively with Dots, introduced on September 29, 2026.

Dots are designed as always-on AI agents capable of pursuing user goals across applications.

Unlike a chatbot that waits for a prompt, Dots are designed to continue working on projects and return with progress, questions or results.

OpenAI says Dots can interact with more than 4,000 applications through its ecosystem.

The agents can also use their own cloud computer and browser to perform tasks.

OpenAI has positioned Dots for both personal and professional workflows, with support for environments such as ChatGPT, Slack and Microsoft Teams.

What Can Dots Do?

Potential applications include:

Research
Software development
Customer support
Sales
Content creation
Data analysis
Project management
Business workflows

The launch shows how quickly AI assistants are evolving from conversational tools into autonomous software workers.

The Big Competition

Dots also put OpenAI directly into competition with Meta's Muse.

The two products represent one of the most interesting AI battles of late 2026.

3. Microsoft Copilot

Microsoft has completely reimagined Copilot.

The redesigned Copilot now revolves around three major experiences:

Home, Code and Autopilot.

Home is designed as a central workspace where users can access conversations, work and Office content.

Code allows users to create software and applications with natural-language instructions.

Autopilot is Microsoft's persistent AI agent designed to continue working on tasks even when the user is not actively interacting with it.

Microsoft is also integrating Word, Excel and PowerPoint directly into the Copilot experience.

That means users can ask Copilot to work with documents, spreadsheets and presentations without constantly switching between applications.

Why It Is Interesting

The new Copilot is less like a chatbot and more like an AI operating layer for work.

Microsoft is clearly trying to make AI part of the everyday workflow rather than a separate tool.

4. Gemini 3.8 Live

Google's AI ecosystem is also moving toward more interactive applications.

One of the newer AI tools highlighted during the September 2026 launch cycle is Gemini 3.8 Live with Live Avatar.

The system combines speech-to-speech AI with near-real-time video generation.

It is designed to create interactive digital avatars that can see, listen and respond naturally.

According to a current roundup of AI launches, Gemini 3.8 Live supports 97 languages and is aimed particularly at business applications.

Why This Could Be Important

Traditional video avatars usually follow predefined scripts.

AI-powered live avatars can potentially respond dynamically.

This could eventually be useful for:

Customer service
Online education
Virtual presenters
Training
Marketing
Digital assistants

The combination of real-time voice and generated video is becoming an increasingly important AI application category.

5. GPT-6 Sol and Luna

OpenAI also introduced new lower-cost AI models during its September product cycle.

GPT-6 Sol and GPT-6 Luna are positioned as more affordable alternatives within OpenAI's newer model family.

Sol is aimed at more demanding professional and coding tasks, while Luna focuses more on high-volume activities such as summarization and data extraction.

Although these are AI models rather than conventional consumer apps, they are important because they can power applications and AI agents.

For users, cheaper high-quality models could mean more affordable AI services.

For developers, they can make it easier to build AI-powered applications without paying the highest inference costs.

6. Grok 4.7

Another major AI development this month is Grok 4.7.

The latest generation is positioned around software development, coding agents and longer professional tasks.

A current September AI launch roundup reports that Grok 4.7 includes a 500,000-token context window and improvements aimed at software development and self-verification.

Who Could Benefit?

Grok 4.7 is particularly interesting for:

Developers
Programmers
Technical researchers
AI enthusiasts
Software teams

The large context window can be useful when working with extensive codebases or lengthy technical documents.

7. Claude Sonnet 5.5

Anthropic's Claude ecosystem remains one of the strongest alternatives in the AI market.

A September 28 update highlighted Claude Sonnet 5.5, which reportedly achieved strong results in Anthropic's terminal coding evaluation while being positioned at a lower token cost than the company's larger Opus model.

This is important because AI coding is becoming one of the most competitive areas in the industry.

Developers increasingly want AI systems that can do more than autocomplete code.

They want systems that can:

Understand entire projects
Run tests
Debug problems
Modify multiple files
Work with terminals
Handle longer development tasks

Sonnet's positioning shows how quickly AI coding assistants are evolving.

8. Granola AI

Not every interesting new app needs to be an autonomous agent.

Granola is an example of a productivity application that focuses on meetings.

Google Play highlighted Granola among its rising apps for September 2026, describing it as an AI meeting-notes application designed to turn conversations into organized notes.

Meeting software has become one of the most practical AI application categories.

Instead of manually writing notes during a meeting, users can let AI help capture:

Key points
Decisions
Action items
Important information
Follow-up tasks

The next generation of meeting assistants is increasingly focused on what happens after the meeting rather than simply producing a transcript.

9. WebRun

AI agents are also becoming more practical for business automation.

WebRun was listed among recent automation product launches in September 2026.

Its positioning is particularly ambitious:

"The AI that runs your company."

The product was launched on September 18 and is part of a growing category of tools designed to automate business workflows.

The broader trend is significant.

Instead of using AI only to generate text, companies are beginning to use AI to coordinate processes.

This could include:

Marketing
Sales
Administration
Customer communication
Internal workflows
Data processing

The category is still young, but it is developing rapidly.

10. Monolayer Autopilot

Another interesting September launch is monolayer Autopilot.

The tool is designed to turn code into self-running software in the cloud.

It was listed as a September 18 automation launch.

This is another example of the growing trend toward software that can operate with less manual intervention.

Developers traditionally have to:

Write code.
Test it.
Deploy it.
Monitor it.
Fix problems.
Update it.

AI-powered automation tools are increasingly attempting to simplify these steps.

11. New AI Tools Are Appearing Every Week

Perhaps the biggest story is not one individual application.

It is the sheer number of new AI products being launched.

A September 2026 launch tracker currently lists dozens of recent AI products across categories including:

Productivity
Developer tools
Marketing
Design
Education
Automation

The latest additions include tools such as Aident Loadout, Astro OS, Fiscus AI, Ricebowl AI, Rixin Archive, Blogwriter and Zanta AI.

This illustrates how quickly the AI software ecosystem is expanding.

Many of these products are still early-stage projects.

Some will disappear.

Others could eventually become major platforms.

What Is Changing in AI Apps in 2026?

The most important trend is the movement from AI assistance to AI action.

A few years ago, the typical AI application looked like this:

User → Prompt → AI → Answer

The emerging model looks more like:

User → Goal → AI Agent → Applications → Actions → Result

That is a much more powerful model.

Instead of generating a response, AI can potentially interact with the software people already use.

AI Agents Are Becoming the New App Layer

The rise of Muse, Dots and Copilot Autopilot suggests that AI agents may eventually become an interface between users and traditional applications.

Consider a small business.

Today, an owner might open:

Shopify
Canva
Slack
QuickBooks
Gmail
Notion
Instagram

and manage each service independently.

An AI agent could eventually become the layer that coordinates these applications.

The user gives the agent a goal.

The agent determines which tools are required.

Then it carries out the appropriate steps.

That is potentially a major change in how people interact with software.

Which New AI App Should You Try?

The best choice depends on what you want to accomplish.

AI App	Best For
Meta Muse	Personal and business AI agents
OpenAI Dots	Autonomous AI projects
Microsoft Copilot	Work and Microsoft 365
Gemini 3.8 Live	Interactive AI and avatars
GPT-6 Sol	Professional AI tasks
GPT-6 Luna	Summaries and data extraction
Grok 4.7	Coding and technical work
Claude Sonnet 5.5	Software development
Granola	AI meeting notes
WebRun	Business automation
Monolayer Autopilot	Cloud software automation

Are These AI Apps Free?

Not necessarily.

One of the biggest changes in the AI industry is that advanced capabilities increasingly come with subscription or usage-based pricing.

Some applications offer free plans.

Others provide limited free access and charge for advanced features.

Business-focused AI agents can also use usage-based pricing because autonomous systems may perform significantly more computing and application actions than a normal chatbot conversation.

For that reason, users should check the current pricing and availability before committing to a service.

What About Privacy?

Privacy should be one of the first things users consider before connecting an AI agent to their applications.

An AI chatbot that only receives a question has relatively limited access.

An AI agent connected to email, financial services, cloud storage and business software is very different.

Before connecting an AI agent, users should check:

What data can it access?
Can it send messages?
Can it make purchases?
Can it modify files?
Can it delete information?
Can it access financial accounts?
Is human approval required?
How is activity logged?
Can access be revoked?

These questions will become increasingly important as AI agents become more autonomous.

The Biggest AI Trend of September 2026

The most important development this month may not be a particular model.

It is the emergence of a new software category:

AI agents that can actually do work.

Meta's Muse, OpenAI's Dots and Microsoft's Autopilot all point toward the same future.

AI is moving from being something users talk to toward something users delegate work to.

That is a significant shift.

Final Verdict

September 2026 has delivered some of the most interesting AI applications and platforms of the year.

Meta Muse is pushing AI agents into personal and small-business workflows.

OpenAI Dots are bringing always-on autonomous agents into the OpenAI ecosystem.

Microsoft Copilot is combining chat, coding and persistent agents inside a single productivity experience.

Meanwhile, newer products are appearing across meeting intelligence, software development, automation, design and business productivity.

The most exciting part is that this market is still developing.

Many of today's newest AI apps are experimental, while others are already becoming useful everyday tools.

The next few months could therefore be even more important.

One thing is increasingly clear:

The best AI app of the future may not be the one that gives you the smartest answer. It may be the one that can take your goal and actually get the work done.