await handleSupport(tickets, { llm: 'local' })

Sarah runs a 12-person SaaS startup. Every morning, her support inbox had 40+ tickets — password resets, billing questions, 'how do I export my data?' She hired a part-time support agent for $2,000 per month. The agent was good at the hard problems, but spent most of her day on repetitive queries that could be answered with a single link. The volume was growing as the user base grew, and hiring another human wasn't the answer. Sarah deployed OpenClaw on a $6 Hetzner VPS on a Friday afternoon. She fed it the company knowledge base, support docs, and her most-common-replies archive. By Monday morning, the system had already auto-resolved 68% of that weekend's tickets. Within a week of tuning, the number rose to 73%, with average response time under 30 seconds — at 3 AM, 3 PM, or any other time. The remaining 27% of tickets that required human attention arrived to her support agent pre-triaged: OpenClaw attached a one-paragraph context summary, the customer's account history, a suggested response draft, and links to the most relevant help articles. What used to take the agent 5 minutes to research took 30 seconds to review and send. Support costs dropped 60%. Customer satisfaction improved from 4.1 to 4.6 out of 5 — because responses were now instant and accurate, even on weekends.