OpenAI has added the ability to perform agentic tasks — that is, sequences of autonomous actions — using voice to the ChatGPT mobile app. The feature is available in the Work tab and is aimed at Pro and Plus subscribers.
For those who manage business processes away from their desks, this means being able to delegate complex operations to ChatGPT by voice, without opening a browser or typing long prompts. It's not an update for everyone: an active subscription is required, and the feature is still limited to the Work tab.
The relevant point isn't the novelty itself, but the direction: AI agents are moving off the desktop and becoming accessible on the go. Those who have already integrated agents into their business workflows need to start thinking about how these tools will change the work of field teams.
What changed in the ChatGPT app, in two sentences
OpenAI has released an update to the ChatGPT mobile app that introduces agentic tasks — sequences of actions that the model performs autonomously, without the user needing to guide it step-by-step — accessible via voice commands. The feature is reserved for Pro and Plus subscribers and is found in the Work tab, already known for advanced productivity features.
It is reported by TechCrunch , which describes accessing via the phone's Work tab as the entry point for completing agentic tasks on the go.
Who really needs it — and who doesn't
The feature makes sense for those who frequently work outside the office: sales managers visiting clients, project managers overseeing construction sites or remote locations, entrepreneurs who use their phone as their primary work tool.
Not needed for those who:
- works mostly at a desk with stable desktop access;
- does not yet have an active Pro or Plus subscription;
- has not defined structured workflows that an agent can execute independently.
The critical point is the latter. A mobile voice agent is only useful if there's something concrete to delegate. Without clear processes behind it, voice becomes a slower way of doing the same things you used to do manually.
The leap this update represents
Until a few months ago, AI agents were desktop tools: configured on screen, launched from a browser, monitored from control panels. With this update, OpenAI is shifting the access point to the phone and making it voice-based.
This isn't an aesthetic detail. It means a purchasing manager can delegate a search to ChatGPT while in the car, or a sales director can start a data analysis while moving between meetings. The interaction model changes: from planned work sessions to micro-delegations spread throughout the day.
Those who have already explored how the OpenAI sponsored agents integrate with platforms like HubSpot and Shopify understands the direction: agents are no longer isolated tools, but nodes in an ecosystem that expands towards new channels and access methods.
Prerequisites and practical limits
Before expecting results, it's helpful to be clear on what is needed:
- Pro or Plus subscription : the feature is not available in the free plan.
- Active Work tab : the access point is specific, not the standard chat.
- Defined workflows : a voice agent performs structured tasks, it doesn't improvise. If the processes haven't been formalized, the result will be disappointing.
OpenAI has not communicated details about additional costs or monthly usage limitations for this specific feature. The available data is limited to what the source reported.
It's also worth remembering that voice introduces a level of ambiguity that written text doesn't have. A poorly formulated voice prompt produces inaccurate output, and in a business context, this can translate into errors on tasks that would require precise instructions. Anyone who has already worked with the Data Agent in the ChatGPT Work tab for data analysis in SMBs you already know the limits related to instruction quality: with voice, that threshold drops even further.
What to evaluate in the coming days
If you have a Pro or Plus subscription, the most useful action is to check if your phone's Work tab already shows the update and test a simple task you normally handle from your desktop.
If you're building or updating workflows with AI agents, now is the time to ask yourself which processes make sense on the go. Not all of them: only those that currently get interrupted because you're not at your desk.
For those managing teams working in the field — sales, technicians, operations — the mobile voice integration opens up scenarios worth exploring alongside other channels. For example, those who have already considered the automated setup with AI agents on WhatsApp Business via MCP you can compare the two approaches: native voice channel vs. structured messaging, depending on the team's operational context.
The AI agent ecosystem expands every week. Following the evolution of automation with AI agents helps you not to miss the moves that really matter for business processes.
Related articles
Discover more articles exploring similar topics, selected to offer you a more complete and stimulating perspective. Each piece of content is carefully chosen to enrich your experience.
