Skip to main content
HomeAboutSolutionsIndustriesContact

AI at Verdeshell

How AI actually works,explained by people who build with it

A working reference rather than a feed. Concepts defined against each other, the patterns we use when we build with agents, and the failure modes each one is guarding against. Where a term is a practitioner coinage rather than a standard, we say so on the page.

Pick a track

Three tracks. Each one is a set of pages that build on each other rather than a stream of posts.

Concepts

The vocabulary — what separates AI from machine learning from generative AI, and an agent from an agentic system.

5 articles

Building with GenAI

How autonomous coding agents are actually structured: loops, graphs, tool access, and knowing when the work is done.

4 articles

Use cases

Applied patterns by function and industry — and an honest account of where the technology does not pay off yet.

In progress

Everything published so far

Loop Engineering: Design the Loop, Not Every PromptA single prompt cannot fix a bug it has never seen run. Loop engineering is the practice of building the harness that lets an agent act, observe, reason and retry until a testable condition is met.6 min readAnatomy of a Good Agent LoopFive things every well-built agent loop needs — and the specific failure each one prevents. If it cannot tell done from not-done, it is not a loop.7 min readGraph Engineering: Coordinating Specialized AgentsWhen one agent is not enough, the question becomes who does what, in what order, and what happens when a step fails. Graph engineering is how that gets specified.6 min readLoops vs Graphs: How the Two Patterns NestOne agent iterating and many agents passing work are different things at different levels. Holding the distinction makes architecture conversations much shorter.5 min readAI, Machine Learning and Generative AI: What the Words MeanFour terms that get used interchangeably and are not interchangeable. They nest — each one is a subset of the one before it, and knowing which you are buying changes what you should expect.6 min readGenerative AI, Agentic AI and AI AgentsThree terms separated by one variable: how much the system decides for itself. Getting them confused is why "agent" projects so often ship as chatbots.6 min readRAG, MCP and Agents Solve Three Different ProblemsThey get compared as if you pick one. You do not — they answer separate questions: what the model knows, what it can reach, and who decides what happens next.7 min readAI Terms, DefinedThe vocabulary you need to read a vendor deck without being misled, grouped by what each term is actually about. Definitions only — no hype.8 min readA Learning Path for Building AI AgentsFive stages, in the order the work actually demands them. Opinionated, deliberately short, and honest about which stages most teams can skip.6 min read

Common questions

What is loop engineering?

Loop engineering describes a single AI agent repeating its own work toward a goal — acting, observing the real result, reasoning about the gap, and retrying until a testable stop condition is met. It is an emerging practitioner term rather than a standardised discipline.

What is graph engineering for AI agents?

Graph engineering describes several specialized agents coordinating as a graph, where each agent is a node with one defined role and each edge defines where work moves next — including feedback edges that route failing work back to an earlier node instead of to a human.

Are loops and graphs alternatives to each other?

No — they nest. A graph governs how work passes between specialized agents; each node in that graph is typically one agent running its own internal loop. In practice you are building a graph of loops.

What does an AI agent loop need to work?

Five things: a goal with a testable termination condition, real tool access so it can act and observe rather than describe, context management for long runs, explicit success and failure logic, and error handling that adapts rather than blindly retries.

What topics does the Verdeshell AI Hub cover?

The hub is organised into 3 tracks: Concepts (The vocabulary — what separates AI from machine learning from generative AI, and an agent from an agentic system.); Building with GenAI (How autonomous coding agents are actually structured: loops, graphs, tool access, and knowing when the work is done.); Use cases (Applied patterns by function and industry — and an honest account of where the technology does not pay off yet.).

Building something with agents?

We build these systems for a living. Bring us the problem and we will tell you honestly whether an agent is the right shape for it.