Find subscriptions and cheaper phone plan to save money
When you buy through links on our articles, Future and its syndication partners may earn a commission. Task: When you buy through links on our articles, Future and its syndication partners may earn a commission. Credit: Shutterstock / Burlingham / Erlin Diah Saving $1,000 sounded pretty good to me. I wasn't entirely sure where Meta's Muse AI was supposed to find it, but Meta's chief AI officer, Alexandr Wang, and the #MuseMoneyChallenge sounded pretty encouraging about putting the company's new AI agent to work finding $1,000 in savings. Some of the results shared so far make that target seem surprisingly plausible. People have reported finding cheaper insurance, forgotten subscriptions, and even unused gift cards, although many of the early examples highlighted around the challenge have come from Meta employees. the last thing your unnecessary expenses see before your muse saves you $1000 pic.twitter.com/NwEO7nhmbK — @alexandr_wang (September 19, 2026) Voir le tweet original I decided to give it a try. I was not expecting Muse to discover a forgotten Swiss bank account, but $1,000 seemed like a sufficiently concrete, if unlikely, target. The results showed how impressive Muse can be, if you're willing to let it into your life. Financial exploration by AI Muse is different from a conventional chatbot because it is designed to take action rather than simply tell you what you should do. Meta says it runs on its own secure virtual computer with a browser, can connect to services such as email and calendar, and can handle tasks including filling out forms and dealing with customer service. It can continue working after you close the app, although actions such as purchases and sending messages can require your approval. That makes the money challenge more interesting than simply asking ChatGPT for tips on saving cash. Telling a chatbot to help me save $1,000 would probably produce an earnest lecture about meal planning and canceling streaming services. Muse can potentially go looking for the forgotten subscription itself. I started broadly and told Muse I wanted it to find ways to save me money. Muse requires a little commitment from the user. The more useful you want an agent like this to be, the more access it needs to the places where useful information lives. Meta lets Muse connect to other apps and services, and says users control those permissions. The company also says login credentials are stored separately so the agent cannot read them, while sensitive actions can require approval. Once I had given it enough to work with, Muse began looking for things that applied to me and surfaced possible actions. I already know that canceling services I do not use saves money. I do not need artificial intelligence to explain the revolutionary concept of paying fewer bills. What I wanted was something able to find the bills I have forgotten about. Muse eventually identified some Patreon and Substack subscriptions I might not want anymore and another opportunity to reduce my phone bill based on some newly released data plans. Between them, the useful savings came to about $35. I checked the suggestions before acting on them, which is something I would strongly recommend doing. Despite saving only 3.5% of the goal of $1,000, I was mildly impressed. It was still $35 I would probably have continued spending every month otherwise. AI is good for boring work Credit: Shutterstock/Poetra.RH The gap between my result and the #MuseMoneyChallenge highlights something important about these increasingly capable personal AI agents. Their usefulness can depend enormously on how much financial debris you have accumulated. The impressive examples circulating online underline the difference in how people may keep an eye on their money, and where AI might actually help those who struggle with doing so. Recovering a year's worth of subscription payments is possible if you have accidentally been paying for something for a year. Those are genuine opportunities, but they are not necessarily waiting in everyone's accounts. My $35 therefore felt less like Muse failing and more like discovering that I apparently did not have $1,000 sitting around waiting for an AI to rescue it. I would also let it spend more time searching for refunds and credits rather than concentrating entirely on recurring expenses.; tools: Meta Muse. Human role: The original post does not say how a person stayed involved.; limitations: Effectiveness depends heavily on the volume of 'financial debris' such as forgotten subscriptions. Meta states credentials are stored separately and sensitive actions require approval, but deep integration with email and billing services is required for the agent to function. Source reliability is secondary, based on a single user experience..
Tools
Published on Agent Work Atlas:
Overview
When you buy through links on our articles, Future and its syndication partners may earn a commission.
Workflow
When you buy through links on our articles, Future and its syndication partners may earn a commission. Credit: Shutterstock / Burlingham / Erlin Diah Saving $1,000 sounded pretty good to me. I wasn't entirely sure where Meta's Muse AI was supposed to find it, but Meta's chief AI officer, Alexandr Wang, and the #MuseMoneyChallenge sounded pretty encouraging about putting the company's new AI agent to work finding $1,000 in savings. Some of the results shared so far make that target seem surprisingly plausible. People have reported finding cheaper insurance, forgotten subscriptions, and even unused gift cards, although many of the early examples highlighted around the challenge have come from Meta employees. the last thing your unnecessary expenses see before your muse saves you $1000 pic.twitter.com/NwEO7nhmbK — @alexandr_wang (September 19, 2026) Voir le tweet original I decided to give it a try. I was not expecting Muse to discover a forgotten Swiss bank account, but $1,000 seemed like a sufficiently concrete, if unlikely, target. The results showed how impressive Muse can be, if you're willing to let it into your life. Financial exploration by AI Muse is different from a conventional chatbot because it is designed to take action rather than simply tell you what you should do. Meta says it runs on its own secure virtual computer with a browser, can connect to services such as email and calendar, and can handle tasks including filling out forms and dealing with customer service. It can continue working after you close the app, although actions such as purchases and sending messages can require your approval. That makes the money challenge more interesting than simply asking ChatGPT for tips on saving cash. Telling a chatbot to help me save $1,000 would probably produce an earnest lecture about meal planning and canceling streaming services. Muse can potentially go looking for the forgotten subscription itself. I started broadly and told Muse I wanted it to find ways to save me money. Muse requires a little commitment from the user. The more useful you want an agent like this to be, the more access it needs to the places where useful information lives. Meta lets Muse connect to other apps and services, and says users control those permissions. The company also says login credentials are stored separately so the agent cannot read them, while sensitive actions can require approval. Once I had given it enough to work with, Muse began looking for things that applied to me and surfaced possible actions. I already know that canceling services I do not use saves money. I do not need artificial intelligence to explain the revolutionary concept of paying fewer bills. What I wanted was something able to find the bills I have forgotten about. Muse eventually identified some Patreon and Substack subscriptions I might not want anymore and another opportunity to reduce my phone bill based on some newly released data plans. Between them, the useful savings came to about $35. I checked the suggestions before acting on them, which is something I would strongly recommend doing. Despite saving only 3.5% of the goal of $1,000, I was mildly impressed. It was still $35 I would probably have continued spending every month otherwise. AI is good for boring work Credit: Shutterstock/Poetra.RH The gap between my result and the #MuseMoneyChallenge highlights something important about these increasingly capable personal AI agents. Their usefulness can depend enormously on how much financial debris you have accumulated. The impressive examples circulating online underline the difference in how people may keep an eye on their money, and where AI might actually help those who struggle with doing so. Recovering a year's worth of subscription payments is possible if you have accidentally been paying for something for a year. Those are genuine opportunities, but they are not necessarily waiting in everyone's accounts. My $35 therefore felt less like Muse failing and more like discovering that I apparently did not have $1,000 sitting around waiting for an AI to rescue it. I would also let it spend more time searching for refunds and credits rather than concentrating entirely on recurring expenses.
Reported result
The author achieved approximately $35 in monthly savings, reaching only 3.5% of the $1,000 challenge goal. The agent successfully identified unused subscriptions and a better phone plan, but the low total reflects the author's existing financial hygiene rather than agent failure. No other quantitative outcomes are reported.
Human role
The original post does not say how a person stayed involved.
Limitations
Effectiveness depends heavily on the volume of 'financial debris' such as forgotten subscriptions. Meta states credentials are stored separately and sensitive actions require approval, but deep integration with email and billing services is required for the agent to function. Source reliability is secondary, based on a single user experience.
Try it
Act as a personal finance assistant using Meta Muse. My goal is to find recurring savings opportunities. [replaceable inputs: Connect to my email and calendar; Target: Identify unused subscriptions (e.g., Patreon, Substack) and compare my current phone plan against available cheaper data plans]. Steps: 1. Scan recent emails for subscription receipts and billing statements. 2. List potential cancellations and plan switches with estimated monthly savings. 3. Present these findings to me for review. Do not execute any cancellations or changes without my explicit confirmation. First, perform a dry run on the last 3 months of email data and show me the top 3 savings opportunities. Wait for my approval before proceeding to any action.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.