How AirOps 5x’d Revenue and Doubled Internal Productivity with Claude

AirOps is a startup building the infrastructure for AI search visibility—helping marketing teams at Ramp, Kayak, Chime, Carta, and Rippling understand how their brand shows up in AI-generated answers and create content to win more citations. The platform does two things: generates long-form content at scale, and evaluates that content against a rigorous quality framework before it publishes.

The core challenge was holding three requirements simultaneously at scale: brand voice, factual accuracy, and SEO performance. Most models degrade on one or more dimensions as prompts grow complex. AirOps scores every piece of content against dozens of checks before publication—paragraph structure, source citations, whether the core answer appears in the first 150 words—and needed a model that could both generate content meeting those standards and evaluate content against them. Early orchestration frameworks were also brittle: setup decisions sometimes required full refactors, and testing new configurations was slow.

AirOps benchmarked models against an internal quality framework and Claude Opus consistently scored highest on the dimensions that matter for marketing content. The team adopted Claude Enterprise org-wide and Claude Code for internal tooling, then migrated to the Claude Agent SDK for the customer-facing agent runtime. AirOps agents operate on “playbooks”—execution instructions that give agents dynamic, adaptive behavior. Claude interprets the playbook, decides which tools to call, and executes across sections rather than following rigid sequences. Sub-agents handle specific subtasks like research, drafting, and evaluation with isolated context windows. The team also built an org-wide skill encoding their full visual and web design system, giving every employee the ability to generate on-brand assets without filing a design ticket.

After adopting the Agent SDK, time-to-quality dropped from 60 hours to 5 hours. Engineering shifted from rerouting LLM calls and updating schemas to refining outputs and catching edge cases before customers hit them. Internally, the consensus across the team was a 2x productivity increase, with individuals reporting 20 to 40 hours saved per week. Customer results were concrete: Chime achieved a 3x citation rate increase with an 89% reduction in content creation time. Carta 4x’d quarterly top-of-funnel output from 5 to 20 articles, with a 75% citation rate on AirOps-built pages. LegalZoom cut Reddit response workflows from 48 hours to under 30 minutes.

AirOps’ agent runtime is now in alpha with customers including Ramp, Carta, Klaviyo, LegalZoom, Xero, Vanta, and AssemblyAI, with general availability planned for mid-2026. The company is building toward full autonomous operation: an agent that identifies an underperforming page, drafts updated content grounded in brand voice, and publishes it—with a human only confirming strategy and making edits when needed. “Claude is central to that vision,” said CEO Alex Halliday.