Zishan Ali Khan builds agentic AI systems for institutions where autonomous failure is a legal event, then publishes the review: field guides, episodes, and the numbers.https://zishanalikhan.com/https://zishanalikhan.com/writing/sovereign-agents/https://zishanalikhan.com/writing/sovereign-agents/If your data cannot leave the country, neither can your agents' reasoning. Notes on the sovereign-AI constraint most multi-agent architectures assume away.Sat, 11 Jul 2026 00:00:00 GMThttps://zishanalikhan.com/writing/agents-skip-tools/https://zishanalikhan.com/writing/agents-skip-tools/An eval finding that changed how I write tool descriptions: agents skip tools because of specs, not capability. The fix is a contract, not a bigger model.Fri, 10 Jul 2026 00:00:00 GMThttps://zishanalikhan.com/guides/agui-field-guide/https://zishanalikhan.com/guides/agui-field-guide/A field guide to AG-UI, the agent-to-user event protocol: event anatomy, transports, process-orchestrated generative UI, and what it leaves unsolved.Fri, 10 Jul 2026 00:00:00 GMThttps://zishanalikhan.com/series/agents-orchestrated/agent-payments/https://zishanalikhan.com/series/agents-orchestrated/agent-payments/AP2 v0.2 plus Camunda 8: signed mandates, a DMN spending policy outside the model, and an audit trail for every hop of an agent-initiated payment.Thu, 09 Jul 2026 00:00:00 GMThttps://zishanalikhan.com/guides/ap2-field-guide/https://zishanalikhan.com/guides/ap2-field-guide/A builder's deep dive into AP2 v0.2: mandate cryptography, the four-party model, threat model, and Human-Not-Present flows, verified against the spec.Thu, 09 Jul 2026 00:00:00 GMThttps://zishanalikhan.com/series/agents-orchestrated/ag-ui/https://zishanalikhan.com/series/agents-orchestrated/ag-ui/AG-UI streamed from a Camunda 8 backend: generative cards, tool calls, and human approvals in one BPMN process, with the audit trail for free.Wed, 08 Jul 2026 00:00:00 GMThttps://zishanalikhan.com/guides/where-agents-work/https://zishanalikhan.com/guides/where-agents-work/Field notes from 50+ enterprise agentic AI builds: where agents earn their keep, by industry and workflow shape, and where plain code still wins.Wed, 24 Jun 2026 00:00:00 GMThttps://zishanalikhan.com/series/agents-orchestrated/loops/https://zishanalikhan.com/series/agents-orchestrated/loops/Every major agentic coding tool converged on the same loop primitive. The moat is not the loop. It is the verification step, and who can audit it.Thu, 18 Jun 2026 00:00:00 GMThttps://zishanalikhan.com/guides/governing-agentic-ai/https://zishanalikhan.com/guides/governing-agentic-ai/Agent frameworks own the inner loop. Your institution owns the outer one. A framework-agnostic pattern for governing agentic AI in production.Wed, 10 Jun 2026 00:00:00 GMThttps://zishanalikhan.com/guides/evaluating-agentic-systems/https://zishanalikhan.com/guides/evaluating-agentic-systems/A field-tested method for evaluating process-embedded AI agents: cost per case, per-field accuracy, tool-skip rates, and temperature sensitivity.Wed, 13 May 2026 00:00:00 GMThttps://zishanalikhan.com/guides/tool-contracts/https://zishanalikhan.com/guides/tool-contracts/Why agents silently skip tools: a failure taxonomy, the spec rules that measurably cut skip rates, and a deterministic 0-100 linter for MCP tool contracts.Wed, 15 Apr 2026 00:00:00 GMThttps://zishanalikhan.com/guides/a2a-field-guide/https://zishanalikhan.com/guides/a2a-field-guide/A builder's field guide to the A2A protocol: Agent Cards, task lifecycle, orchestration vs swarm, privacy boundaries, and the threat model that matters.Wed, 04 Mar 2026 00:00:00 GMT