Ramp shipped invoicing this week, a notable expansion since they’ve been a spend-side product for years. The product itself is fine. The UX worth studying is the send flow, which is three screens: pick a customer (autocomplete from existing contacts or paste a name and email), build the line items (single inline-editable table, no modals), preview-and-send (the rendered invoice plus a one-line message). That’s it. No “set up your first invoice template” wizard, no “configure your tax settings” prerequisite, no “verify your business address” flow. They make defensible defaults for all of that and let you fix them later.
Compare to Bill.com’s invoice flow (8 screens, 3 modals) or QuickBooks (which still wants you to map a chart of accounts before you can send a single bill). The implicit thesis: “the first invoice should send in under 60 seconds; configuration is a tax we charge users who keep using us.” What we’d flag: the preview screen renders the invoice client-side as PDF using a slightly older PDF.js build, and the page jitters as it composes. Server-side render the preview, send the bytes, done. Pattern to steal: when porting a flow from a desktop accounting tool to a web product, the right mental model is “what does a defensible default look like?” not “how do we collect every preference up front?” You can always add settings later. You can’t undo a 17-step onboarding.