Since its launch in September, Meta’s Muse, an artificial intelligence (AI) assistant designed to perform tasks on behalf of users, has quickly become the most downloaded app in its category. In response, OpenAI, the developer behind ChatGPT, introduced a rival AI assistant named dots, available through a $100-a-month subscription. Both programs represent a new wave of AI "agents" — digital tools capable of managing routine activities such as booking appointments, applying for jobs, and organizing daily schedules by effectively interacting with other apps and web services.

Unlike traditional chatbots that primarily offer direct answers, these AI agents operate like autonomous digital workers with their own web browsers and access to various applications. This functionality enables them to complete more complex assignments, a capability that earlier virtual assistants from companies like Amazon, Apple, and Google struggled to deliver to widespread consumer satisfaction due to technological limitations.

Anthropic, another AI developer, recently announced a similar project called Duet, signaling an emerging race to refine AI agents capable of handling everyday tasks more effectively.

Testing both Muse and dots over a two-week period revealed the potential impact of such agents on simplifying routine responsibilities:

1. **Restaurant Reservations**: At Sun Moon Studio, a Michelin-starred restaurant in Oakland, California, with only a dozen seats, securing a weekend noon booking is notoriously difficult. The AI agents were instructed to attempt reservations repeatedly. While initially unsuccessful, dots eventually confirmed a booking for late November, underscoring the agents’ ability to monitor and capitalize on openings in real-time.

2. **Workout Planning**: When asked to generate a workout plan upon arriving at a local climbing gym, Muse succeeded by sending a tailored routine immediately upon detecting the user's arrival. Dots could not perform this task as it lacks location-triggered actions. This illustrates potential uses tied to context-aware scheduling or task reminders.

3. **Job Applications**: For assisting a former nanny in applying for caregiver positions, Muse demonstrated effectiveness by autonomously responding to numerous job postings on relevant websites, even producing positive responses from potential employers. Dots encountered technical glitches during the process, indicating that the technology remains imperfect.

4. **Meal Planning**: Both assistants were tasked with suggesting daily lunch and dinner ideas based on dietary preferences and available ingredients. Each morning, the agents provided tailored menus, helping relieve the mental load of family meal preparation.

5. **Price Negotiation**: When searching for tires, Muse and dots identified an online deal significantly lower than the local tire shop’s price and successfully submitted price-matching requests to the retailer, resulting in confirmed price matches via email within hours.

Users setting up these AI agents must carefully consider the permissions granted, as accessing personal emails and other apps involves sharing sensitive information, which companies may review to improve their services.

Alexander Embericros, head of enterprise product at OpenAI, noted that dots was initially intended for more complex business management tasks, with plans to eventually offer the service free to users after initial paid testing.

As AI agents like Muse and dots continue to advance, experts suggest they may soon become commonplace tools for automating repetitive tasks, shifting how users manage daily activities and freeing time for higher-value endeavors.