AI FIELD MANUAL
Not a hundred ChatGPT prompts and not a listicle. Each entry starts with one specific, realistic situation and walks through how an AI practitioner would actually solve it - the approach, the reasoning, the exact steps, the real tools, and where the method breaks down.
A specific situation, not a category - a role, a deadline, a real friction. If the problem could apply to anyone, it isn’t specific enough to act on.
The recommended workflow, and the actual reasoning behind it - not just what to do, but why this method rather than the more obvious shortcut.
Every tool is real, checkable, and never paid placement. Every entry names where the method breaks down - a workflow with no stated limitations isn’t finished.
A realistic run of the workflow end to end, so you can see the shape of the result before you spend the time.
Also in Nakoda Intelligence
AI EncyclopediaEvery technique named in these workflows, defined precisely and sourced.
Decisions that need to survive a board room, not a demo.
Workflow
How to build a defensible AI-competitive-landscape brief for a board meeting without an LLM quietly inventing a competitor's product, partnership or funding round.
Small team, real deadline, no room for a wrong workflow.
Workflow
A structured, AI-assisted way to turn a stack of raw customer discovery transcripts into honest themes, using a real qualitative-research method rather than asking a chatbot to "summarize the feedback."
Pipeline
A real evaluation methodology - golden dataset, LLM-as-judge with a written rubric, and a regression gate - for teams about to ship an AI feature, instead of shipping on "it looked right in testing."
Workflow
A source-first market-sizing method - AI structures the framework, but every figure on the slide traces to a named, checkable source, never a model's confident-sounding estimate.
Where the model meets the codebase, the logs and the on-call rotation.
Pipeline
A concrete retrieval-augmented workflow for using an LLM to debug a production incident against your actual codebase and logs - not against its training data's vague idea of what your framework does.
Pipeline
A real evaluation methodology - golden dataset, LLM-as-judge with a written rubric, and a regression gate - for teams about to ship an AI feature, instead of shipping on "it looked right in testing."
Evidence discipline - what a claim needs before it goes in a paper.
Pipeline
Why asking an LLM to "write a literature review with citations" produces fabricated references, and the retrieve-then-summarize workflow that avoids it.
Using the tool to learn the material, not to skip it.
Pipeline
Why asking an LLM to "write a literature review with citations" produces fabricated references, and the retrieve-then-summarize workflow that avoids it.
Checklist
A daily study workflow that uses AI for Socratic questioning and spaced-repetition card generation - active recall - instead of as an answer machine, plus where the line to academic dishonesty actually sits.
Volume and speed, without the brand voice or the facts drifting.
Workflow
A real content-gap workflow: export competitor keyword data from an SEO tool, then use an LLM only to cluster and prioritize it - never to invent the ranking data itself.
Also in Nakoda Intelligence
AI EncyclopediaEvery technique named in these workflows, defined precisely and sourced.
Exploration and iteration speed, with a real handoff at the end.
Workflow
Using AI UI-generation tools for fast, constraint-driven concept exploration - then treating the output as a starting point for real design work in Figma, not a finished deliverable.
Repeatable process, with an audit trail that survives a review.
Pipeline
A concrete retrieval-augmented workflow for using an LLM to debug a production incident against your actual codebase and logs - not against its training data's vague idea of what your framework does.
Decision tree
A decision-tree approach to AI-assisted workflow automation that routes the routine cases automatically and keeps a human, and a paper trail, on every exception.
Decision tree
A confidence-threshold approach to AI-assisted invoice data extraction and reconciliation that speeds up the routine matches and routes anything uncertain to a human - never auto-posts on a guess.
Numbers that reconcile, with a human accountable for the exceptions.
Decision tree
A decision-tree approach to AI-assisted workflow automation that routes the routine cases automatically and keeps a human, and a paper trail, on every exception.
Decision tree
A confidence-threshold approach to AI-assisted invoice data extraction and reconciliation that speeds up the routine matches and routes anything uncertain to a human - never auto-posts on a guess.
Judgment about people, with the documentation a regulator can read.
Checklist
Why free-form 'rank these resumes' AI screening is a legal and fairness risk, and the rubric-first alternative that keeps a human accountable for every decision.
A deliverable a client can act on, sourced well enough to defend.
Workflow
How to build a defensible AI-competitive-landscape brief for a board meeting without an LLM quietly inventing a competitor's product, partnership or funding round.
Workflow
A source-first market-sizing method - AI structures the framework, but every figure on the slide traces to a named, checkable source, never a model's confident-sounding estimate.
Getting a good answer to a real question, and knowing when not to trust it.
Checklist
A verification-first checklist for using AI to plan a complex, multi-country trip - great for structure and ideas, dangerous for any fact you'd actually act on without checking.
Nakoda’s AI Strategy division works with teams to design and implement workflows like the ones in this Field Manual, tuned to a specific business.
AI Strategy ↗For entries touching compliance, hiring, or finance controls, AI Audit and AI Governance review the workflow before it goes live.
AI Governance ↗The tools named in this Field Manual change fast. The Nakoda newsroom tracks what actually shipped, not just what was announced.
UAE AI News ↗
