Agentic features
Also written: Autonomous AI, AI agency, Agent features, AI agents
AI that can act autonomously to perform tasks, interact with tools, and make decisions.
Agentic features mean AI can do things on its own. Instead of just answering a question, it can take action. This could involve using other apps or making a series of choices to complete a larger goal.
For example, imagine telling your AI assistant to "plan a weekend trip to the mountains." An agentic AI could then research destinations, check weather forecasts, find available hotels, and even book a room, all without you needing to give step-by-step instructions for each part.
Agentic features refer to AI systems designed to operate autonomously or semi-autonomously in performing tasks. This involves capabilities such as planning, decision-making, tool use, and interaction with digital or physical environments. These systems can perceive their surroundings, reason about goals, and execute actions to achieve them, often iteratively.
Key aspects include the ability to break down complex goals into sub-tasks, select and utilize appropriate tools (like APIs or software interfaces), and adapt to changing conditions. Tradeoffs involve increased complexity in development and potential risks related to unintended actions, necessitating robust safety mechanisms.
Recent updates from OpenAI and Google Gemini highlight AI agents that can perform multi-step tasks and interact with other applications. This focus on AI performing actions autonomously, rather than just responding to prompts, is a key development in making AI more useful in daily tasks.
- ChatGPT mobile app gets voice-based agentic featuresOpenAI · TechCrunch AI ·
- GitHub Copilot weekly releases — September 14Microsoft · GitHub Changelog: Copilot ·
- Moving to the new ChatGPT desktop appOpenAI · OpenAI Help Center ·
- Ask Gemini to handle your multi-step tasks in select Android appsGoogle · Gemini Help Center ·
- Ep 836: Updated GPT-5.6, A new Cheap Meta Model, Qwen 3.8 released and impressive and 7 more AI updates you can use todayOpenAI · Everyday AI ·
- Wayfair boosts catalog accuracy and support speed with OpenAIOpenAI · OpenAI News ·
- The Kleenex of AIGoogle · AI Inside ·
- Google Gemini
- Google Gemini is enabling AI agents to automate multi-step tasks within select Android applications by interacting with the screen.
- OpenAI ChatGPT
- OpenAI is integrating agentic features into its ChatGPT mobile app, allowing users to perform complex tasks using voice commands.
- Task automation
- Task automation typically involves pre-programmed scripts or ru
- Chatbots
- Chatbots primarily focus on conversational interaction, while agentic
- Robotic Process Automation (RPA)
- RPA automates structured, rule-based tasks, of
The concept of agents in AI has a long history, dating back to early AI research. The modern focus on agentic features in large language models has been significantly shaped by research papers and community discussions around enabling LLMs to interact with the world and perform complex tasks. Source ↗
Open a tool like Google Gemini or ChatGPT. Ask it to perform a multi-step task that requires using another application or service, such as "Find the weather in London tomorrow and then draft an email to a friend suggesting we meet up if it's nice."
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- WorkflowBuild an AI agent evaluation pipeline in one hourCompress weeks of AI agent testing into one hour using open-source tools and automated pipelines. · 60 min
- HeuristicAI agents need time-awareness for runtime controlAI agents require a sense of time to predict, estimate, and control their own actions and execution duration. · 5 min
- Anti-patternUsing prototype defaults for production AI agent architectureDon't let prototype settings for AI models and prompts become production defaults; they are not optimized for scale or cost. · 60 min
- Cost tipLeverage prompt caching to reduce AI agent token costsReuse previously processed prompt prefixes to avoid reprocessing and save on token costs for agents. · 30 min
- HeuristicComposable controls for agent autonomy and safetyUse composable controls that perform single jobs, enabling agents to run with autonomy at appropriate safety levels. · 10 min
Questions people ask
- What's the difference between an AI assistant and an agentic AI?
- A traditional AI assistant primarily responds to direct commands. An agentic AI can take initiative, plan, and execute a series of actions to achieve a goal with less direct supervision.
- Are agentic features safe?
- Agentic features introduce new safety considerations because the AI can act autonomously. Developers are working on safeguards to ensure these actions are aligned with user intent and do not cause harm.
First seen on Subvolts Mar 11, 2026. Updated Oct 5, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.