Practical AI writing for people who have to use it.
Comparison guides, how-to playbooks, concept definitions and field essays for teams turning AI from hype into measurable workflow change.
Insights & Resources is the writing arm of Applied AI North, a Toronto practice offering AI consulting, AI literacy training, and web development for Canadian businesses. It covers human-in-the-loop systems your team actually uses and owns, written by practitioners who ship, for the people who have to make it stick.
Published
Two pieces so far, newest first.
-
Many AI projects stall around month seven. Not because the model regressed, the API moved, or the team lost interest, but because the question of who pays the cost of the system working was never named at month zero. A short diagnostic for AI projects in production.
-
Most AI “agents” are one of three things in a trench coat. Naming them correctly is the first decision that affects scope, price, and evaluation strategy. A field taxonomy with examples.
Forthcoming
What we’re writing next, grouped by topic. None of these are published yet.
Comparison guides
Long, structured comparisons of the tools, models, and patterns we use daily. Each one ends with a recommendation, not a shrug. Updated whenever a new model release moves the verdict.
-
n8n vs Make vs Zapier for AI workflows
When to reach for which workflow tool when LLM steps are part of the equation. Self-hosting, branching logic, vendor lock-in, and the breakpoint where a workflow tool stops being the right answer.
-
LangGraph vs Vercel AI SDK vs raw OpenAI Agents
Three approaches to building a production agent loop. Each one optimizes for something different: control, ergonomics, or velocity. The one we reach for most often, and why.
How-tos & playbooks
Written as numbered steps with the assumption that you will actually try them. Pulled from real engagements, with the parts that don’t work crossed out.
-
How to write your first eval set
A 9-step guide to building the labeled test set that should exist before you write a single prompt. Includes the templates we use, the trap most teams fall into at step 4, and the size you actually need (smaller than you think).
-
A 6-hour AI literacy curriculum for non-technical teams
The exact agenda we run for ops, sales, legal, and finance groups: what these tools actually do in 20 minutes, where they fail, 3 workflows you’ll use on Monday, and the privacy material nobody asks about until lunch. Slides included.
-
12 cost-control levers for production AI agents
Caching, prompt compression, model routing, batch inference, context truncation, semantic dedup, fallback ladders, and 5 more. Each lever with the typical savings, the trade-off, and the line of code that turns it on.
-
Confidence floors: the math, the bug, and the fix
When to set the floor at 0.80, when to set it at 1.00, and the subtle bug that bit us on a production extraction agent for half a week. Worked example with code. The right floor depends on the cost of a wrong answer versus the cost of a human review.
-
A 3-week build playbook for AI-aware websites
The week-by-week schedule we use to ship custom websites at a fraction of typical agency cost: where AI compresses the work, where human judgment stays in the loop, and the 3 places where speed should never be the priority.
Field essays
The essays we write when something repeats often enough across engagements to be worth naming. No fixed schedule. No newsletter funnel, no listicle quotas. The first, The Six-Month Rule, is published.
-
Evals are the deliverable
The reason we write the test set before the agent. A field guide to building eval harnesses that actually steer the project, plus the 3 failure modes of eval sets that look correct on paper.
-
Adoption is the project
A taxonomy of 3 failure modes that kill AI rollouts after launch: the pilot purgatory, the leadership-line gap, and the trust deficit. What to instrument to catch each one early.
-
The case against discovery phases
Why every 14-week discovery we’ve seen could have been a 2-week assessment and a courageous “no.” What clients actually need at the start of a project, and what they’re being sold instead.
-
Stop calling it a copilot
The word has lost any specific meaning. A more useful three-way taxonomy (concierge, reference, operator) that changes how scope, evals, and pricing land in the same conversation.
-
What a real handoff looks like
The 5 artifacts every AI engagement should produce, the test for whether the handoff actually transferred ownership, and the one question that exposes a fake one in 30 seconds.
-
Context is the moat, not the prompt
Why the durable advantage in applied AI is the quality of context you can assemble at the moment of a request, not the cleverness of the wording. 3 exercises to find where your context is leaking quality.
Canada & compliance
A growing part of this index, because this is where we work. Plain-English writing on Canada’s AI policy, the Voluntary Code of Conduct, the federal Guide on Agentic AI, data sovereignty, and the funding programs that may help with the cost of adoption. We write it to translate the “AI for All” strategy into practical steps for real systems, with conditional framing and program terms you should verify before you act.
-
PIPEDA, AIDA, and the Canadian SMB in 2026
What actually governs SMB AI projects today (PIPEDA, Quebec Law 25, and the Voluntary Code of Conduct) now that the proposed AIDA lapsed with Bill C-27 (Jan 2025) and new federal rules are being shaped under the “AI for All” strategy. The 3 steps to take this quarter, regardless of how the legislation lands.
-
Data residency for Canadian AI: a vendor checklist
Which model providers offer Canadian or US-only data-residency, what to ask for in a data-processing agreement, and the small set of patterns that lets sensitive client data stay inside your own infrastructure.
-
A plain-English guide to Canada’s Voluntary Code of Conduct
The six elements of the Voluntary Code of Conduct on the responsible development of advanced generative AI (Accountability, Safety, Fairness, Transparency, Human Oversight, and Validity & Robustness) read as a checklist you can align a real system to, not a press release.
-
Bounded Autonomy: the federal Guide on Agentic AI, explained
The four autonomy levels in the federal Guide on the Use of Agentic AI, why we build Level 1–2 assistive systems with read-only access by default, and how recoverability (a literal kill switch and immutable logs) defends against automation drift and prompt injection.
-
Which Canadian programs actually fund AI adoption?
A working map of what exists and who each program suits: BDC LIFT financing, the Ontario Job Grant (training from accredited providers only), the Productivity Super-Deduction (capital investment, your accountant’s call), and FedDev Ontario’s AI adoption funding for larger projects. Each is subject to current program terms.
-
Data sovereignty: keeping AI workflows on Canadian soil
How we architect for Canadian data residency: deployable on Canadian cloud regions (e.g. AWS Canada Central) and Canadian foundation models such as Cohere, so proprietary data and IP can stay onshore, engineered to align with PIPEDA, Quebec Law 25, and AIDA readiness.
A working glossary
Short definitions of the terms we actually use in client conversations, written so a non-technical operator can use them in a meeting on the same day. Longer explainers for each entry are forthcoming.
What is applied AI?
Applied AI is the use of language models, retrieval, and agentic workflows to change how specific work actually gets done inside a business: as a production system that measurably saves time or adds throughput, not as a demo or a chatbot. The emphasis is on adoption by the people doing the work.
What is context engineering?
Context engineering is the deliberate design of what information a model sees at the moment of a request: system prompt, retrieved documents, tool descriptions, history, examples. It accounts for more of output quality than prompt wording does. Most production AI failures are context failures, not prompt failures.
What is an eval harness?
An eval harness is the code and data that lets you measure an AI system’s quality against a labeled test set, automatically, on every change. It is the closest thing AI engineering has to a unit-test suite. Without one, every change is a guess and every regression is a surprise.
What is human-in-the-loop?
Human-in-the-loop (HITL) is a system design where an AI produces a draft and a human reviews, edits, or approves it before the outcome is committed. Where a step may run on its own, it does so only above a confidence floor you agreed was safe; below it, a person decides. HITL is how production AI stays both fast and trusted.
What is anchor prompting?
Anchor prompting is a pattern where the full instruction set lives in an uploaded reference document, and chat prompts invoke it by name. This stabilizes long agent sessions, reduces context rot, and makes instructions easier to version and update than inline prompts.
What is context rot?
Context rot is the degradation of model output quality over a long conversation as the context window fills with irrelevant or contradictory information. Symptoms include rule-following drift, repeated answers, and ignored constraints. The fix is structured truncation, not a bigger window.
What is a confidence floor?
A confidence floor is the model-reported probability threshold below which a response is routed to a human reviewer instead of being committed automatically. Typical values are 0.80–0.95. The right floor depends on the cost of a wrong answer relative to the cost of a human review.
What is AI literacy?
AI literacy is a working understanding of what current AI tools can and cannot do, how to use them effectively for a specific job, and how to recognize when their output is wrong. It is closer to a research skill than a technical one, and it is the highest-impact training a non-technical team can do this year.
Questions we keep answering
Short, direct answers to the questions that come up on the first call with new clients. Each answer is kept short enough to stand on its own. Same questions, longer answers, live across the rest of this page.
-
What is applied AI?
Applied AI is the use of large language models, retrieval systems, and agentic workflows to change how specific work actually gets done inside a business: not as a demo or a chatbot, but as a system in production that measurably saves time, reduces errors, or adds new throughput.
-
How long does a typical AI project take?
Most applied-AI engagements run 4 to 10 weeks end to end. A 2-week AI Roadmap scopes the work; a fixed-price build of 2 to 8 weeks (one workflow, or one assistant) delivers a working system, hardens it and hands it over to the team. Projects that run much longer usually have an organizational adoption problem, not a technical one.
-
How much does an AI engagement cost?
Published prices at Applied AI North run from CA$750 a month for AI Office Hours to CA$15,000 for the largest fixed-price build. The 2-week AI Roadmap is CA$2,500, a workflow automation is CA$2,500–9,000, and an AI assistant is CA$7,500–15,000. Workshops start at CA$1,500. Illustratively, two or three engagements combined come to about CA$10,000–18,500 over two to four months.
-
What is the difference between an agent and a chatbot?
A chatbot responds to messages. An agent decides what to do, executes tool calls, observes the results, and loops until a task is complete. Agents have memory, planning steps, and side-effects: they send emails, write to databases, or call APIs. Chatbots produce text; agents produce outcomes.
-
How do you measure AI return on investment?
Measure the workflow the AI replaces, not the AI itself. Track time-per-task before and after, error rates against a labeled ground-truth set, human-review percentage, and adoption rate (daily active users among the intended audience). A 92% accurate system used by the whole team beats a 99% accurate system nobody trusts.
-
What is context engineering?
Context engineering is the deliberate design of what information an AI model sees at the moment of a request: the system prompt, retrieved documents, tool descriptions, conversation history, and structured examples. It accounts for more of the output quality than prompt wording does. Most production AI failures are context failures, not prompt failures.
-
What is an eval set?
An eval set is a labeled collection of input-output examples used to measure how well an AI system performs against your real data, before and after every change. For a document-extraction agent, that means 100 to 500 historical documents with the correct answers attached. The eval set is the deliverable; the agent is the side-effect of building it.
-
Are AI projects different for Canadian businesses?
Yes. We build systems engineered to align with the frameworks actually in force today (PIPEDA, Quebec Law 25, and Canada’s Voluntary Code of Conduct on generative AI) and built for readiness for the federal AI regulation signalled under the 2026 “AI for All” strategy. The proposed Artificial Intelligence and Data Act (AIDA) was part of Bill C-27, which died at prorogation in January 2025; there is no federal AI statute in force right now. Practically: we architect for Canadian data residency where your data calls for it, keep agentic systems to bounded, human-in-the-loop autonomy, and itemize invoices so your accountant can judge any tax treatment.
-
What is human-in-the-loop?
Human-in-the-loop (HITL) is a system design where an AI generates a draft or decision but a human reviews, approves, or edits it before the outcome is committed. Where a step may run on its own, it does so only above a confidence threshold you agreed was safe; below it, a person decides. HITL is how production AI stays accurate and trusted.
-
How do I know if my team is ready to adopt AI?
Three signals: leaders can describe the workflow they want to change in plain English; one named person on the team is willing to own the system after launch; and you can produce 20 real examples of the work the AI will do. If any of these is missing, start with a literacy workshop before a build.
New pieces by email. No tracking, no funnel.
RSS or email, your call. Plain text. We will never sell your address, and we will not send anything that isn’t writing.
Or use the RSS feed: /feed.xml
Reading with a project in mind? See what we do: AI Consulting AI Literacy & Enablement Web Development