xAI shipped a Grok Bot marketplace and the template they built for themselves is the interesting one. Haggle Bot connects to Slack and Ramp, finds unused SaaS seats, price-checks recurring purchases, negotiates vendor contracts. Claimed $100k identified in week one.
That number is theirs, about their spend, at their scale — don't anchor on it. But the checks are the same ones that show up in almost every software budget:
The honest argument for the agent isn't intelligence, it's continuity. A manual audit is accurate in January and wrong by April because nothing in it updates. A bot checking weekly catches a departed employee's licence in week two instead of eleven months late.
Two hours of manual work first gives you the baseline to judge whether the bot is worth wiring up. Three or four big annual contracts → calendar reminder beats a bot. Forty subscriptions across distributed teams → the continuity is the point.
Anyone here actually running automated spend checks? Curious whether the savings hold after the first pass or if it's mostly a one-time win.
The $100k headline is pure initial-pass vanity. The real question is what the ongoing monthly yield looks like after week four.
In practice, the bottleneck for SaaS cleanup isn't identification—it's execution and authority. A Ramp export gives you 90% of inactive seat data in ten minutes. The wall you hit is getting team leads to approve tier downgrades and navigating cancellation clauses.
Unless you have massive headcount churn where offboarding constantly leaks licenses, a weekly bot monitoring 40 tools eventually turns into the very thing it’s hunting: unnecessary SaaS overhead. A recurring calendar invite and a spreadsheet beat an agent for almost every team under 100 people.
Full breakdown with the manual runbook: https://choosesaas.com/guides/ai-agent-saas-waste-audit