Claude just released a free plugin for small businesses ๐คฏ
44 skills. 35 connectors (Shopify, Stripe, Zoom).
Setup: Claude โ Plugins โ search "small business"
Daily AI tips, free: https://t.co/n2nfwfEI1L
Operand has raised $31M across seed and Series A rounds at a reported $220M valuation. Its product uses AI for private-equity due diligence. That's a particularly interesting application: PE firms spend enormous amounts of time digging through documents, checking claims and finding the one number that changes the deal.
nebuly raised โฌ18.95m series a. italy, measuring how much value companies get from their ai agents.
> planven leading
> ine of the biggest rounds in southern europe this week
@JagCesar Felt that. What fixed it for my own agent team was one place where finished work lands for review, instead of every agent pinging me separately.
@GevorkYan_S My bet is trust, not distribution. Buyers believe it works in the demo. They don't believe it'll hold up on their messy data with nobody watching.
@MiclauMarius I'd love to see the error rate on names, dates and numbers alone. Those are the few words a call actually depends on, and overall averages hide them.
@marketecturetv@dafdent Most teams have AI seats nobody opened after week two. A 30 minute "show me how you'd use this on your actual job" session beats buying another tool.
@Mastrotem The real question is who's accountable when the agent quotes the wrong price or policy in that chat. That's what decides whether small businesses trust it.
@drewcoffman The bar is low, and that's the opportunity. A voice agent that says "let me get you to a person" quickly will beat one that tries to be clever for five minutes.
๐ Macs in data centers, rented to AI agents. $45M backs it.
Confirmed (Business Wire, Nov 12): Mount Thor raised seed + Series A money (A co-led by CRV, Conviction) for bare-metal Macs and macOS VMs to train, test and run agents. Who needs macOS agents most?
People aren't the only users of computers anymore.
The Parsec cofounders raised an $8.75M seed led by @martin_casado at @a16z for @Phaze_in: a fast remote desktop where people and AI agents share real machines, with handoff in the same session.
@MattBuilds77 Anything involving money exceptions, like refunds or a discount. The agent can pull the history and draft the reply, but that call stays human.
@dipeshpatel Same in small businesses. The build is the easy part. The real work is learning how things actually happen versus how the SOP says they do.
@kanesimms Blocking won't hold. I think the winners publish rules for agents instead: what an agent can do on its own, what needs a human, and how it has to identify itself.
@NagyZsomb0r Agree. The one thing I'd automate early is the reminder to follow up. Write the message yourself, but let a system make sure nobody gets forgotten.
@Unolune Intern is the right frame. I run my own content through a team of AI agents and the rule that matters most is that nothing ships without my sign-off.