March 26, 2026
When AIs do the sales call
Show HN: Claude skill that evaluates B2B vendors by talking to their AI agents
Dev builds AI that grills sales reps so you don’t have to — founders are cheering, salespeople are fuming
TLDR: A new tool lets an AI interrogate software vendors’ own chatbots and public info to rate who’s actually a good fit, instead of relying on slick sales pitches. Commenters are split between loving the protection from bad deals and worrying it’s just two robots confidently misleading each other.
Hackers just met their new office hero: an AI that interviews software vendors’ own chatbots, then tells you who’s bluffing. The creator says it’ll save teams from endless sales calls and glossy pitch decks, but the real show is in the comments, where engineers, founders, and sales pros are going to war.
On one side, burned buyers are loving it. Commenters shared horror stories of being sweet-talked into year‑long contracts for tools that “basically turned into a fancy spreadsheet.” For them, an AI that cross‑checks vendor claims against review sites and public data is pure revenge. Several joked this is like having a “robot procurement goblin” that lives only to say no to sales reps.
Sales and vendor folks, though, are nervous and snarky. Some grumbled that it’s just “outsourcing due diligence to yet another black box,” while others worried that feeding everything through vendor chatbots just means two AIs lying to each other. The meme of the day: screenshots imagining a vendor AI and buyer AI locked in an endless loop of “circling back” and “per my last message.”
Underneath the jokes, there’s a sharper debate: are we finally getting honest software comparisons, or just inventing a new way to automate bias? The thread reads like couples therapy for the entire software industry, and nobody’s leaving unscathed
Key Points
- •Buyer Eval is a Claude skill that automates structured, evidence-based evaluations of B2B software vendors for buyers.
- •The skill researches the buyer’s company automatically, asks domain-specific questions, and sets hard constraints such as budget, compliance, and integrations.
- •It engages vendor AI agents through the Salespeak Frontdoor API and cross-checks their claims against independent sources like G2, Gartner, and LinkedIn.
- •Vendors are scored across seven dimensions with transparent indication of vendor-verified versus public-only evidence, and results include TL;DR, scorecards, risk analysis, and demo prep questions.
- •The skill is installed via Git clone from GitHub, supports both global and per-project installs, auto-checks for updates, and can be invoked in Claude Code or Claude desktop with the /buyer-eval command.