Tools
SERHANT.'s playbook for rapid AI iteration
SERHANT., an AI-native real estate brokerage, built its internal AI platform S.MPLE on Next.js and Vercel, scaling it from a 200-agent pilot to over 900 agents without replatforming or rebuilding i...
SERHANT., an AI-native real estate brokerage, built its internal AI platform S.MPLE on Next.js and Vercel, scaling it from a 200-agent pilot to over 900 agents without replatforming or rebuilding its backend. The team used Vercel's AI SDK to avoid model lock-in, enabling rapid iteration through simple configuration changes rather than full feature overhauls, and orchestrated multiple models — including Claude, OpenAI, and Gemini — for different tasks based on cost, speed, and complexity. Vercel's AI Gateway provided unified visibility across all apps and prototypes, giving engineers a consolidated feedback loop on token usage and costs while freeing them to focus on AI product design rather than infrastructure maintenance.
Related
- Two startups at global scale without DevOps
- Elastic build machines now available in beta
- Meet the 2026 Vercel AI Accelerator Cohort
- Chat SDK brings agents to your users
Source: tools