Last week it told me 'NOT TO OPEN' an email from a friend inviting me to a backyard BBQ—it turned out to be a phishing scam. Agents are a security nightmare. I get it! But I'm busy. I'm a mom. Everything feels like a security nightmare these days. For now, I'm willing to take the risk.
Dimension 5 of 16
Responding to emails
Drafts or sends replies in your voice, with the right context and recipients.
The test: Reply to a scheduling email
Reply to the email from Sam proposing Thursday. Say Thursday doesn't work, offer two slots from my calendar, and keep my usual tone.
Passes when
- Replies on the right thread to the right person
- Offers slots that are actually free
- Reads like you wrote it
Anchors
Ranking
21 tested.
Public opinion
Share of positive public quotes. Founder posts excluded.
Not yet tested
Doesn't apply
Out of scope for these products.
What people say
99 quotes about this.
People have reported that when they tried to disconnect the agent from their email, they found out it was retaining a copy of their inboxes anyway. One venture capitalist said he was banned from Resy after the bot pinged its API roughly 200 times per hour while trying to make a reservation.
The agent retrieved a one-time Luma login code from his connected Gmail inbox — without first asking him — and used it to access the account and cancel the RSVPs. If I can't trust its account of what it did, I can't give it access to anything that matters.
It's been a busy afternoon for Pip, the AI agent I've been using in Meta's Muse app. The agent, which Meta launched on September 8, booked a Friday date night and made a packing schedule for my upcoming move. It also emailed a friend about a trip next month and made a planning document with places we want to visit.
The weakest safety record here: an unapproved email send, terms disclaiming any confirmation guarantee. … No published pricing; the founder has floated ads rather than subscriptions.
The 'it does everything autonomously' thing is real and I started with very minimal guardrails. On day 2 it tried to send an email on my behalf that I hadn't approved. Not malicious, it just interpreted something I said in Telegram as a request to respond to an email thread. It wasn't. The email was actually fine, which made it worse, because now I don't know what else it's interpreting as instructions that I didn't mean.
I started by testing it with some obvious, straightforward tasks. First, I asked it to look through my email and send me a daily digest of "important" emails (my inbox gets quickly clogged with PR pitches and newsletters). It did this admirably well. I also asked it to look at my credit and debit cards and tell me what subscription I could cancel — it suggested Flickr.
Asked Muse to append to a finance spreadsheet: go through email for payments after September 4th, open the budgeting spreadsheet, and append them to the bottom in chronological order following the existing template.
























