How to Use AI to Build a World's Economic Underbelly

Every fictional world has an official story it tells about itself — the tax records, the merchant guild charters, the royal mint, the regulated trade routes. That story is almost always a lie. Or at least, it's incomplete in the ways that matter most for fiction. The real texture of a world lives in what happens when the official story breaks down: the back-alley surgeon who works without a license, the favor owed to a smuggler who got your family across the border, the dried herbs that aren't technically illegal but everyone knows what they're for.

Shadow economies aren't just set dressing. They're the connective tissue between your power structures and your street-level characters. When you write a book with AI, it's easy to end up with a world that feels like a Wikipedia article — lots of correct information, zero lived friction. The economy section describes the currency, the major exports, maybe a trade war. But nobody's haggling in a flooded basement, nobody's paying off a debt in stolen grain, nobody's laundering money through a traveling puppet troupe. This guide is about building that layer.

Why Official Economies Lie and Shadow Economies Tell the Truth

Here's the thing about official economies in fiction: they exist at the wrong altitude for storytelling. A character interacts with the Royal Merchant's Guild once, maybe twice, and probably only when something goes catastrophically wrong. But they interact with the informal economy every single day — bribing the gate guard with a meat pie, trading gossip for lodging, getting a cut-price healing potion from someone who definitely didn't acquire it through official channels.

Shadow economies are where economic pressure becomes personal. A tariff is an abstraction. A debt to the woman who controls all the unlicensed river crossings in your region — that's a plot. The more you understand what your world's laws and institutions can't or won't provide, the better you understand where the informal economy fills the gap. And those gaps are almost always where your most interesting characters live.

This is true across genres. Fantasy has its underground magic markets and its smuggled artifacts. Sci-fi has gray-market tech, black-site organ replacement, data laundering. Post-apocalyptic fiction basically is informal economics. Even contemporary thrillers live in the spaces between official transaction records. If you're working on fantasy worldbuilding specifically, the economic underbelly is often where your magic system and your class system collide in the most interesting ways — who can afford healing, who sells secrets, who controls the unofficial distribution of power.

The unofficial economy doesn't exist despite your world's laws. It exists because of them. Every prohibition creates a market. Every shortage creates a workaround. Every official barrier creates someone whose entire livelihood is helping people around it.

Keep that principle in your back pocket. It's the engine for everything that follows.

Mapping the Gaps: Prompting AI to Find What Your World's Laws Can't Cover

Before you can build the shadow economy, you need to know where the shadows are. This is actually one of the places AI is most useful — not generating content from scratch, but stress-testing a system you've sketched and finding its holes.

The prompt pattern here is what I'd call the prohibition audit. You give the AI your world's official rules and ask it to identify what those rules make impossible, scarce, or dangerous — and who would therefore profit from providing it anyway. This is different from asking "what does my black market sell?" That question gets you a list. The prohibition audit gets you a logic.

Why this works: you're not asking for generic fantasy content. You're giving the AI a specific institutional structure (theocratic monopoly on healing), specific populations with specific needs (non-believers, working class, criminals, travelers), and asking for the logical consequences. The output won't be a random list of black market items — it'll be a system with internal cause and effect. Tweak the institution and the populations to match your world, and you'll get something genuinely tailored.

Do this exercise for every major institution in your world that controls access to something people need. Legal systems create illegal legal aid. Food rationing creates food black markets. Immigration controls create people-smugglers. Magic licensing creates unlicensed practitioners. Work through your world's institutions one by one, and by the end you'll have a map of where the shadows fall.

This also sets you up well when you later need to edit a book with AI — because the AI will have a coherent economic logic to reference rather than a patchwork of invented details that contradict each other mid-manuscript.

From Smuggling Routes to Favor Debt: Building Transactional Systems That Drive Plot

Once you know what's being traded informally, you need to build the how. Shadow economies aren't just black markets with price tags. They run on different transactional logic than official economies — and that logic is what creates story.

Favor debt is the most underused economic mechanism in fiction. In an informal economy, favors are currency. They're also more dangerous than money, because they're non-specific — when someone calls in a favor, the debtor doesn't always know the price until the moment of payment. That's inherent narrative tension, and it costs you nothing to build it into your world's economic logic from the start.

Here's where AI can help you think in systems rather than individual transactions. Ask it to design a transactional logic — not just what's traded, but how trust is established, how debts are tracked, how disputes are resolved, and what happens when someone defaults. These mechanisms will naturally generate plot-relevant situations.

That last line — "give me enough detail that I could write a scene without consulting anything else" — is a useful general principle when prompting for worldbuilding. It forces the AI toward operational specificity rather than high-level summary. You want to know that disputes are settled by an Elder Council that meets on the third Wednesday of each month in the hydroponics bay, not that "disputes are handled through community mediation."

Smuggling routes deserve the same treatment. A route isn't just a line on a map — it's a series of handoffs, each one involving a different person with their own motivations and risks. If your character needs to move something illegal from point A to point B, the interesting question isn't whether they succeed. It's who they owe favors to afterward, who knows what they moved, and which of those people will show up later wanting something. Consider building your routes using the same approach described in guides like How to Use Conditional Prompts to Write Branching Scenes — because a smuggling run is essentially a branching scene with economic stakes at each decision node.

Prompt Examples: Getting AI to Invent Specific Black Markets, Currencies, and Trade Slang

Generic worldbuilding feels generic because it's been processed through too many levels of abstraction. "There's a black market for magical artifacts" is an abstraction. "The Tarnished Stair on Rope-Seller's Row, where you can fence a minor enchantment for three tallow-coins or trade it for two doses of unlicensed forgetting-smoke" is a place. The specificity is what makes it real.

AI is very good at generating specificity when you ask for it directly. The mistake most writers make is prompting at the abstraction level and then being disappointed by abstract output. Push past the category into the granular.

The grounding instruction — "make everything feel like it emerged from the whale-iron industry" — is doing a lot of work here. It forces specificity and prevents the AI from defaulting to generic fantasy market tropes. Your world's informal economy should smell like your world. It should use the byproducts and the failures and the excesses of your official economy as its raw material.

For trade slang specifically, I find it helpful to ask for slang in context — not just a glossary, but example sentences showing how a character would actually use the terms. That way you can hear whether it sounds natural or whether it reads like an anthropologist's field notes.

Tone instruction ("cynical and street-level," "unglamorous work") prevents the AI from drifting toward the theatrical. Shadow economy workers in real life — and in good fiction — treat their work with the same weary pragmatism as anyone else doing a job. The drama comes from the stakes, not from characters constantly remarking on how shadowy their dealings are.

If you're building these systems for a series rather than a standalone, you'll want all of this captured in a central reference. The Series Bible Template is worth using here — economic systems are exactly the kind of detail that drifts inconsistently across books if you don't lock them down early. Pair it with a Proper Noun Extractor to catch your trade slang and market names so they stay consistent through the manuscript.

Stress-Testing Your Economy: Using AI to Find the Contradictions Before Readers Do

Here's a scenario that happens to almost every worldbuilder: you've built a detailed shadow economy, you're proud of it, and then someone in your beta readers points out that your black market for healing potions couldn't exist because you established in chapter two that healing magic requires a specific mineral that the Church controls the only mine of. The whole system collapses under scrutiny.

Readers who love your genre will stress-test your world whether you want them to or not. Better to do it yourself first, with AI as your adversarial reader. This is one of the most practically useful things AI can do for fiction writers — not generate content, but find the holes in content you've already made.

The adversarial audit prompt pattern is simple: you give the AI your economic system in as much detail as you have, and you ask it to find contradictions, impossibilities, and unintended consequences. You're not looking for the AI to fix the problems — you're looking for it to surface them so you can make deliberate choices about how to address them.

"Don't soften your analysis. I need the real problems, not reassurance." This instruction matters. AI models have a tendency to be encouraging about creative work, which is actively unhelpful when you're in diagnostic mode. You need the problems. Ask explicitly for them.

This kind of structural review pairs naturally with a full manuscript edit. When you're ready to edit a book with AI, your economic systems will hold up much better if you've already done this adversarial pass. The Five-Pass Revision Order for AI-Assisted Novels is worth reading alongside this — worldbuilding consistency is exactly the kind of thing that benefits from a dedicated pass rather than trying to catch everything at once.

One final thing worth stress-testing: character consistency within your economic system. If you've established that your protagonist grew up in the shadow economy and spent ten years working as an informal courier, their instincts around transactions, trust, and risk should reflect that. An character consistency check can catch places where your characters behave like outsiders to systems they should know intimately. Economic fluency is characterization. Getting it wrong is as damaging as getting their dialogue voice wrong.

Also consider: what happens to your shadow economy under pressure? What if there's a crackdown, a famine, a war? Healthy worldbuilding means knowing not just how the system works in equilibrium, but how it fails. Ask AI to model a specific disruption — "the Church of the Pale Hand suddenly floods the market with free healing" — and trace the downstream effects on your informal economy. Those disruptions are often where your best plots live.

If you're still building the scaffolding for your world and want a structured starting point, an AI book outline can help you figure out which parts of your shadow economy need to be load-bearing for your plot and which are just flavor — useful to know before you've spent three sessions generating black market slang for a district your protagonist never visits. For deeper worldbuilding infrastructure, How to Build a Story Bible That AI Models Actually Follow covers how to document these systems in a way that AI actually uses consistently rather than ignoring after chapter four.

The practical takeaway: before your next writing session, pick one institution in your world — legal system, food distribution, medical care, whatever — and run the prohibition audit prompt on it. Ask AI what that institution makes impossible, scarce, or dangerous, and who profits from providing it anyway. One institution, one hour, one session. You'll come out with more story material than you know what to do with, and it'll all be rooted in the specific logic of your world rather than imported from generic fantasy tropes. That's the difference between a world that feels built and a world that feels lived in.