Field notes
Clear thinking for teams making consequential technical decisions.
Original guidance on the architecture, controls, and operating models behind enterprise AI and SaaS products.
Latest Zarafah insights
Agentic systems · 7 min read
What changes when an AI agent becomes a production system
A practical framework for moving from an impressive agent demo to a controlled, observable business capability.
SaaS engineering · 8 min readSaaS architecture decisions that matter more in AI products
How tenancy, usage, permissions, evaluation, and model flexibility reshape the SaaS platform foundation.
Knowledge systems · 6 min readEnterprise RAG reliability starts before retrieval
Why source ownership, permissions, structure, freshness, and evaluation determine whether enterprise retrieval can be trusted.
Have a specific system question?
Bring us the context.
A useful conversation starts with the decision, constraint, or failure mode your team is working through.