ANALYSIS10 min · 2026-07-08
Gartner's 2025 agentic AI cancellation forecast: questions to review.
A forecast is not an individual diagnosis. Review ownership, costs, tests, expectations, and use-case fit against the project's evidence.
Field notes
Implementation playbooks from the field.
Production AI patterns we ship every week. Stuck pilots, MCP hardening, eval suites, agent orchestration, cost control. Written by Alex Cinovoj, Founder & CTO, TechTide AI · CTO & Co-Founder, CPG Playbook. No vendor pitches, no recycled blog posts.
10 of 71 field notes · 2026 archive
All topicsAgentsCoding agentsCost controlDeliveryEvalsGovernanceHuman in the loopMCPModel choiceObservabilityOrchestrationRetrievalROISecurity
July 2026
EXPLAINER9 min · 2026-07-01
MCP vs A2A: two protocols, two different jobs.
MCP is vertical, model to tools. A2A is horizontal, agent to agent. Choose based on whether you need tool access or coordination between independently operated agents. The decision rule, with failure modes.
June 2026
ANALYSIS10 min · 2026-06-17
The harness matters more than the model now.
Model swaps stopped moving the needle. The gains now come from the harness around the model: durable state, checkpoints, retry policy, budget ceilings, and clean recovery.
FIELD NOTE8 min · 2026-06-10
When to hand an agent back to a human.
The agent does not need to be right every time. It needs to know when it is out of depth. Four escalation triggers, plus the handoff payload that makes a human fast.
May 2026
SECURITY9 min · 2026-05-13
Prompt injection is an access-control problem.
Filtering instructions out of untrusted text is a losing game. The defenses that hold are boring: least-privilege tools, typed arguments, allowlists, and approval on writes.
ANALYSIS10 min · 2026-05-06
The only three orchestration patterns that survive contact with production.
Swarms demo well and page you at 2am. Three orchestration patterns survive real traffic, and the choice is decided by failure isolation, not elegance.
April 2026
PLAYBOOK9 min · 2026-04-22
If you cannot replay it, you cannot ship it.
Debugging an agent from a chat transcript is guesswork. Log the full run as a trace, keep cost and latency on the same timeline, and alert on tool-level failures.
FIELD NOTE8 min · 2026-04-08
Autonomy is a dial, not a switch.
Teams argue about whether the agent should be autonomous. Wrong unit. Set autonomy per action, using reversibility and blast radius, and earn each step up with evidence.
March 2026
ANALYSIS9 min · 2026-03-11
You probably do not need the framework.
Fourteen nodes in a state graph to answer refund questions. Size the stack to the problem: one loop, typed tools, a log, and only then a framework.
January 2026
EXPLAINER10 min · 2026-01-28
Context engineering is the job now.
The prompt is the smallest part. Production quality comes from context assembly: retrieval scope, ordering, compaction, tool results, and a hard token budget.
More from Alex on buying and engineering
Buying guides on alexcinovoj.com and engineering notes on alexcinovoj.dev, published alongside these field notes.
Companion reading
- Buying guide · alexcinovoj.com
What drives the cost of an AI implementation?
How to read an implementation quote: process clarity, data preparation, integrations, review and recovery, and operating ownership. A proposed evaluation method, not a price list.
Published October 29, 2026 · January 2026 collection
- Buying guide · alexcinovoj.com
Buy software or commission an AI integration?
A decision guide for choosing between software you already own, an extension and a custom integration, with exit paths and a bounded trial before expanding. The examples are hypothetical.
Published October 29, 2026 · February 2026 collection
- Buying guide · alexcinovoj.com
How to choose an AI implementation partner
A procurement method that asks every supplier the same questions about scope, working evidence, acceptance criteria, security boundaries and handoff. It does not rank agencies.
Published October 29, 2026 · February 2026 collection
- Engineering note · alexcinovoj.dev
Inside ClawKeeper: testing permission boundaries before an agent acts
A walkthrough of ClawKeeper's deterministic agent policy, tenant checks and payment approval gate, based on seven synthetic policy tests run at one inspected revision, with its known trust boundaries.
Published October 29, 2026 · January–March 2026 collection
- Engineering note · alexcinovoj.dev
Inside CipherClaw: how recorded agent traces become an investigation map
A source-code reading of CipherClaw's trace graph, root-cause heuristics and breakpoints at one inspected revision, with its implementation limits. The review did not run the test suite.
Published October 29, 2026 · January–March 2026 collection