Tech • IA • Crypto
How AirOps chases friction to build AI products with Claude
6/10TL;DR
AirOps is redesigning AI agent building for marketers by replacing complex workflows with document-based systems, improving usability, speed, and output quality.
KEY POINTS
Shift from workflow builders to agents
AirOps began with node-based workflow builders similar to automation tools, allowing users to chain steps for content creation. These systems proved too complex for non-technical users, especially marketers unfamiliar with concepts like JSON or templating. As AI models evolved rapidly, workflows became brittle, requiring constant updates and creating maintenance burdens.
Targeting marketers in AI search
The company focuses on “AI search,” where brands optimize visibility in tools like ChatGPT, Claude, and Gemini. Its platform helps businesses analyze how they appear in AI-generated answers, identify gaps, generate or update content, and measure performance improvements such as citation rates and share of voice.
Launch of AirOps Next and Quill agent
A new release introduced Quill, an AI agent designed as a “content strategist” for marketers. Quill integrates brand data, search insights, and content guidelines, enabling users to move directly from identifying gaps to executing content strategies without complex setup.
Document-based “playbooks” replace technical builders
AirOps replaced traditional workflows with playbooks, a document-style interface resembling familiar tools like Google Docs. These act as reusable “skills” where users define inputs, outputs, and tools in natural language. The approach lowers the barrier to entry while preserving structure, versioning, and collaboration.
Human-in-the-loop governance
A central design focus is embedding human review throughout the workflow. Users can approve, edit, or comment at specific stages, acting as gatekeepers before content progresses. This ensures brand alignment and reduces the risk of low-quality or incorrect AI-generated material.
Transparency and control in agent behavior
Despite simplifying the interface, the system exposes which tools are used and how context is applied at each step. This maintains user trust and allows marketers to understand and control outputs without needing engineering expertise.
Specialized tools improve efficiency
Instead of relying on repeated generic tool calls, AirOps built specialized tools that perform complex analyses in a single step. For example, a page analysis tool can instantly evaluate SEO gaps, competitor comparisons, and keyword opportunities, reducing token usage by 8% and significantly improving speed.
Use of sub-agents for quality and focus
The platform employs multiple sub-agents with specific roles, such as compliance checks, writing, and brand context retrieval. This modular approach prevents context overload and improves output consistency by isolating tasks within focused environments.
Enterprise results and adoption
Early enterprise users reported strong gains, including a 130% increase in citation rate and 42% growth in share of voice. Time to deployment dropped from about a month to as little as one week, with teams able to self-serve rather than rely on technical support.
Balancing flexibility with intentionality
A key challenge identified is the overwhelming number of potential use cases for AI agents. AirOps emphasizes narrowing focus to specific workflows, such as structured content creation processes, to avoid sprawl and maintain product clarity.
Ongoing challenges: feedback loops and benchmarking
Future work includes improving agent self-learning through better feedback systems and developing benchmarks for content quality. Unlike coding tasks, content evaluation is subjective, making it harder to measure whether system changes genuinely improve outcomes.
CONCLUSION
AirOps’ approach highlights a broader shift in AI product design toward simplicity, governance, and domain-specific tooling, aiming to make powerful agents usable for non-technical professionals while maintaining enterprise-grade quality.