CENDAR FIELD NOTES
Practical guides to AI search, smarter workflows and the systems behind them. Find an idea you can apply to your website, your team or your next project.
Make a useful page findable, readable and measurable.
Check the task, evidence and limits before choosing a model.
Understand the software around the model and how to compare it.
Give each action an owner, permission boundary and recovery path.
Audit AI search visibility with a URL evidence log, a real Cendar Lab example and a blank worksheet. Separate access, indexing, citations and visits.
Read the guide : AI Search Visibility Audit: Find What’s Holding Your Website Back
A sourced briefing on Jev, GPT-6 Sol and Luna, and Claude Fable and Mythos 5.1, with practical checks before adopting a model.
Read the guide : AI Models in September 2026: What’s Worth Evaluating for Your Business?
TypeSafe AI introduced Jev on September 15, 2026. Here is what its Choice, Score and Noul outputs can do, and what still needs evaluation before automation.
Read the guide : Jev by TypeSafe AI: Where Could It Fit in Your Workflow?
A source-based comparison of three coding agents: what the harness controls, which boundaries to inspect and how to run a fair task evaluation.
Read the guide : Codex vs Claude Code vs Pi: Which Coding Harness Fits Your Team?
An agent's permission prompt, process sandbox, container and microVM each protect a different boundary. Start with the code and data the task can reach.
Read the guide : Sandbox vs MicroVM: Choose the Right Boundary for Your AI Agents
Use the model card, license, task fit, hardware needs and a local evaluation to narrow candidates on the Hugging Face Hub.
Read the guide : Choosing a Hugging Face Model? Start with These Deployment Checks
Orchestrate agents around clear task ownership, tool permissions and checked outputs. Add parallel work only when dependencies allow it.
Read the guide : AI Agent Orchestration: Turn Separate Tools into a Working Process
Repeated long prompts can increase latency and cost. Learn when prefix caching helps, what to compact, and which measurements to record before changing an agent workflow.
Read the guide : Prompt Caching and Context Compaction: Where to Look for LLM Cost Savings
Combine vector and keyword retrieval for RAG. Learn where PostgreSQL full-text search differs from BM25, how RRF works, and what to evaluate before deployment.
Read the guide : Hybrid Search with pgvector: Full-Text Search, BM25 and RRF
Answer Engine Optimization helps make useful information discoverable, understandable and supportable in AI answers. For a B2B company, it builds on SEO rather than replacing it.
Read the guide : What Is AEO? A Practical Guide to Getting Your Business Ready for AI Search
Training, search indexing and user-initiated fetching are three different requests, and each vendor gives them a different user agent. Decide them separately in robots.txt instead of allowing or blocking “AI” as one thing.
Read the guide : AI Crawlers: Which Bots Should Your Business Website Allow?
Google states that no AI text file is needed to appear in its AI features, and the llms.txt proposal never claimed to be a ranking signal. It describes an on-demand file for agents. Decide whether you have that use case before adding one.
Read the guide : Does llms.txt Help AI Search? Where It Fits in Your Website Strategy
Structured data can describe visible page content, but it does not guarantee an AI citation. Learn when schema markup helps and how to validate what you publish.
Read the guide : Structured Data for AI Search: Make Your Business Easier to Understand
Overlapping pages can confuse readers and maintenance. Learn when to merge content, how to preserve useful material and how to verify redirects and internal links.
Read the guide : Content Consolidation: Give Competing Pages a Clearer Purpose
What can Search Console show about AI search features? Separate documented reporting from estimates, then pair query data with the referrals you can actually observe.
Read the guide : AI Search Traffic in Google Search Console: What Should You Track?
When several pages make different claims about the same company, which should a reader trust? Review source ownership, factual consistency and visible evidence before adding more markup.
Read the guide : Build a Consistent Business Identity Across Your AI Search Content
Break down speech recognition, model response, speech output and telephony delay. Learn where to measure, test interruptions and keep tool results from becoming premature promises.
Read the guide : Voice AI Latency: Find the Delays That Interrupt the Conversation
MCP standardizes how hosts connect to tools and context servers. Learn what the protocol covers, what the application still owns, and which controls to test before deployment.
Read the guide : MCP in Production: Connect AI Agents to the Tools Your Business UsesCompare a managed model API with dedicated inference using your prompt volume, concurrency, latency target, hardware costs and operating responsibilities.
Read the guide : Local vs Cloud LLM Inference: Which Cost Model Fits Your Workload?
GEO began as an academic paper on optimizing content for generative engines. AEO is a practitioner term for being usable as a direct answer. SEO is the discipline both build on. The differences that matter are practical, not terminological.
Read the guide : AEO vs GEO vs SEO: Where Should Your Business Focus?
Measure AI citations with a defined question sample, recorded test conditions and source-level review. Keep mentions, supporting citations, referral visits and business outcomes separate.
Read the guide : How to Measure AI Citations and Decide What to Improve Next
Review the real page, its claims, discovery paths and next steps before adding more content. This checklist produces an evidence-backed improvement backlog, not a score that promises AI citations.
Read the guide : Is Your Website Ready for AI Search? Use This AEO ChecklistAn AI agent harness is the software around a model that manages context, tool execution, and the loop between actions and results.
Read the guide : What Is an AI Agent Harness? The Software That Turns a Model into a WorkflowDefine the trigger, field ownership, permissions, and failure cases before asking for a connector or a quote.
Read the guide : CRM Integration: What to Define Before You Connect Your ToolsCheck what your current tools can already do. Build custom code for a specific gap, not simply because an API exists.
Read the guide : Native Connector vs Custom Integration: Choose the Right Fit for Your WorkflowSeparate routine record creation from the decisions that need a person, including scope changes, missing information, and access approval.
Read the guide : Client Onboarding Automation: Move New Customers from Sale to DeliveryCalendar access, appointment status, privacy, and human review must be defined before an intake assistant handles real requests.
Read the guide : AI Appointment Intake: What to Check Before Automating BookingsAnswering a question, moving a record, and deciding which tool to call are different jobs. Start with the one you need to improve.
Read the guide : Chatbot vs Workflow vs AI Agent: What Does Your Business Actually Need?Build an inbound process around permission, a responsible person, and a next action. Stop follow-ups when the conversation changes.
Read the guide : Lead Follow-Up Automation: Keep the Next Sales Conversation Moving