Intake agent for a 4M-customer bank
Triages support tickets, drafts replies, routes escalations to the right team. Shipped in 5 weeks. 86% precision on the gold set, $0.04/run inference cost.
Names real where we have permission. Numbers honest. Every case study links to a sample of the deliverable — the SOW, the eval set, the report — wherever we can share it.
Triages support tickets, drafts replies, routes escalations to the right team. Shipped in 5 weeks. 86% precision on the gold set, $0.04/run inference cost.
New cluster strategy, technical SEO fixes, 60-page content sprint over 90 days. Migrated off a JS-rendered stack that was blocking indexation.
Weekly LinkedIn + X cadence built around the CEO's voice. 4 short-form videos/week from existing podcast footage. DM-ops handled by our team.
Rebuilt the Meta + Google funnels, server-side tracking, new LPs in our stack, full attribution wired to their CRM. Killed non-performing channels.
AEO-first content rebuild across the top 40 pages. llms.txt, schema rollout, citation outreach to the sources AI engines pull from.
Internal portal for evidence collection, audit prep, and policy management. TypeScript + Astro + Postgres, deployed in their VPC.
Copilot that answers questions about a customer's own contract library. Built in their codebase, runs in their VPC with existing auth.
Migration from a legacy CMS to a headless stack we built. Pre-migration audit, redirect map for 4,800 URLs, post-launch monitoring.
A line per engagement. Ping us if you want the long form on any of them.
Tell us what you're trying to ship. We'll tell you what we'd do, whether it fits, and what it would cost.