Cyril blog https://getcyril.com/blog/ How Cyril is built — an AI-native business operating system. en What to ask an AI vendor about permissions https://getcyril.com/blog/what-to-ask-about-ai-permissions/ https://getcyril.com/blog/what-to-ask-about-ai-permissions/ Eight questions that decide whether an AI feature can be turned on for everyone or only for a pilot group — with what a good answer sounds like. Sam Akbari Mon, 31 Aug 2026 00:00:00 GMT What changes when your development team is an AI https://getcyril.com/blog/when-your-development-team-is-an-ai/ https://getcyril.com/blog/when-your-development-team-is-an-ai/ Cyril is built by AI, directed by one person. What that changes about writing software — and why it doubles as a test of the product's own claim. Sam Akbari Fri, 28 Aug 2026 00:00:00 GMT The integration tax: what five separate tools really cost https://getcyril.com/blog/the-integration-tax/ https://getcyril.com/blog/the-integration-tax/ An open, assumption-by-assumption model of what five disconnected business systems cost a 40-person company — and why licences are the smallest line. Sam Akbari Tue, 25 Aug 2026 00:00:00 GMT The seven-question test for whether AI is really built into a platform https://getcyril.com/blog/seven-question-test-for-ai-native/ https://getcyril.com/blog/seven-question-test-for-ai-native/ Seven questions that tell you whether a platform's AI is architectural or a feature line — with the strong and the weak answer to each one. Sam Akbari Tue, 18 Aug 2026 00:00:00 GMT Letting AI act safely needs one audit log, not five https://getcyril.com/blog/agents-need-one-audit-log/ https://getcyril.com/blog/agents-need-one-audit-log/ Generative AI answers questions; agentic AI takes actions — and actions are where trust is lost. Why safety is one permission model and one audit log. Sam Akbari Tue, 02 Jun 2026 00:00:00 GMT Every record explains itself: the ai_context serializer https://getcyril.com/blog/every-record-explains-itself/ https://getcyril.com/blog/every-record-explains-itself/ Every Cyril entity has an ai_context serializer: a deterministic, schema-aware view built for the model rather than the database. Why grounding needs it. Sam Akbari Wed, 20 May 2026 00:00:00 GMT Why AI-native platforms win — and what that actually means https://getcyril.com/blog/why-ai-native-platforms-win/ https://getcyril.com/blog/why-ai-native-platforms-win/ Bolted-on AI demos well and falls apart in real workflows, for structural reasons. The engineering decision that separates AI-native from AI-added-on. Sam Akbari Fri, 08 May 2026 00:00:00 GMT