How to use AI to build smarter TCGplayer pricing rules

“Stay competitive.” “Protect the expensive stuff.” “Move bulk.” Those are reasonable goals, but each one needs boundaries. Which game, finish, and condition? How far can a price fall? Ask an assistant to compare the rules with your actual listings. You still choose the tradeoff.

My sequence is read → propose → review. First, look at current rules, listings, movers, and sales. Then decide which groups need their own rule. Only then ask for a proposal and inspect it in Hoard before approving it. “Fix my prices” skips the part where you decide what the rule should do.

Start with a decision, not a multiplier

Start with the job each group of cards needs to do. The multiplier comes later:

• Protected inventory: preserve margin or avoid selling scarce, high-value, collector, or volatile cards too cheaply.
• Competitive inventory: stay reasonably positioned on repeatable cards where turnover matters.
• Clearance inventory: move slow or low-value stock, but only above a floor you are willing to accept.

Those goals collide if one rule covers everything. A collector foil shouldn't inherit the bulk rule just because both are Magic cards. A Lightly Played copy may need a different ladder from Near Mint. Ask: “Where is one rule making two different decisions for me?”

The read-first prompts

Keep the first pass read-only. A connected Hoard assistant can check inventory, rules, pricing explanations, store-wide diagnostics, movers, sales, and sync health without changing the store.

1. Find rule collisions and missing lanes

Read my active pricing rules and diagnose my store. Which rules cover the most cards, which cards have blockers or stale matches, and where do my games, finishes, or conditions appear to be sharing a rule that may not fit both? Do not propose changes yet.

Look for one rule covering more cards than you expected, or a rule you thought was active that is disabled or expired. Then inspect a few sample cards. The assistant can compare the saved price explanation with a current-rule test, so you can see whether a lock, hold, stale match, or different rule is involved. Use the store-wide diagnosis guide for a broad pattern and the one-card guide for a single listing.

2. Separate a game before comparing its behavior

For Magic, Pokémon, and One Piece separately, summarize my current listed value, active-rule coverage, recent movers, and cards that sold in the last 30 days. Flag material differences, but do not assume generic market norms are true for my store.

Your sales and listings may show a different price mix or turnover pattern for each game. They can't prove a sweeping claim such as “all Lorcana commons should cost X.” For newer games, make the assistant show your store data instead of relying on its general card knowledge.

3. Test finish and condition as independent decisions

Within my Magic inventory, compare regular cards, foils, and other variant treatments. Then compare NM, LP, MP, HP, and Damaged where I have enough listings. Show current rule coverage, listed value, recent sales, and a few representative cards per segment. Read-only; do not draft a rule yet.

You don't need a rule for every finish or condition. You do need to know when one shared rule is making a choice you wouldn't make yourself. See condition-tier price ladders and foil and sealed-product pricing for two common places this happens.

Set a floor before you ask for a competitive rule

“Follow the market” still needs a stopping point. A thin market or a single cheap listing may not be a price you're willing to take. Set your minimum in Hoard's pricing settings before asking for a more competitive rule. The assistant's current rule proposal cannot set that account-level number for you.

Then make the assistant show its work:

For one sample card in each proposed segment—bulk, regular singles, foils, and high-value cards—show the floor and protection layers in the current-rule test. Read my active rules and never-go-down settings. If you cannot see my account-wide numeric floor, tell me to confirm it in Hoard Settings; do not invent a value. Do not create a proposal.

Hoard checks proposals on the server, but it can't know an acquisition cost you haven't recorded or the price at which you'd rather hold a card. Decide that first. The floor and ceiling guide and fee-aware repricing guide help you set the boundary.

Turn the decision into a proposal—not a price write

Once the evidence and floors are clear, ask for a bounded rule matrix. Be explicit about the scope and what must remain protected:

Prepare a pricing-rule proposal only. Keep my current floors and protections. Create separate rules for (1) Magic regular cards priced competitively, (2) Magic foils and premium treatments with a more protective strategy, and (3) low-value Magic inventory with a clearance strategy. Flag locked cards and any segment where the available data is insufficient instead of assuming they will reprice. Before I approve it, summarize every rule’s scope, before/after settings, matched-card count, listed-value impact, and protected-card warnings.

Hoard handles pricing-rule changes as propose → review → apply. The assistant prepares a proposal; the review page shows each rule's before/after settings, matched-card counts, listed-value impact, and protected-card warnings. Nothing in that proposal applies until you approve it. If a targeted rule changed meanwhile or a safety check fails, the whole change set stops and you can ask for a fresh proposal. The rule-matrix walkthrough covers the clicks.

Approval changes rules; it does not give the assistant permission to set any card price it likes. Hoard applies the approved proposal itself. Future repricing still follows your pricing setup and safeguards. To see how a strategy behaves before live repricing, use watch-only mode and rollback as a separate check.

How to review affected cards like an owner

Don't approve a rule because its card count looks plausible. Open the proposal and check:

1. Scope: Does each rule include the intended game, finish, condition, and price tier—and exclude the rest? A rule meant for ordinary Magic singles should not quietly capture premium treatments or another game.
2. Representative cards: Open a handful of cards from each lane: a cheap card, a middle card, a high-value card, and any card the proposal warns is protected. Ask why each is governed by that rule and inspect the price breakdown.
3. Outliers: Sort attention toward the largest expected impact, unusual percentage moves, thin or missing market data, price locks, and cards with a stale match. Do not accept a broad result because the average looks reasonable.
4. Business consequence: Listed-value impact is exposure, not revenue. A higher total is not a sale. Compare the result with your floor, turnover goal, and reason for making that lane.

A useful final prompt is: Before I review this proposal, list the affected-card segments, the largest expected changes, all protected-card warnings, and any assumptions you made because the data was incomplete. Do not apply it. It gives you a short owner’s checklist without pretending that the assistant can make the judgment for you.

Know the limits before you trust the answer

AI can read and reason over the data Hoard exposes. It cannot turn incomplete evidence into certainty, know the condition of a card it has not inspected, or reliably supply market lore for every game. Treat broad claims about unfamiliar games, variants, or collector demand as hypotheses unless they are supported by your data or a source you independently trust.

The write boundary depends on the action. An assistant can propose a pricing-rule matrix for review, but manual per-card overrides, locks, rollbacks, CSV import/export, new inventory, and account settings remain in the dashboard. Other actions, including some Smart Actions, can apply directly if you grant write access. Keep the first audit read-only and check the current capability map before asking for a write.

A useful assistant can show what it read, which cards a rule would touch, and what it couldn't know. You decide whether that rule belongs in your store.

Related documentation: What your AI assistant can do · connect an AI assistant to your store · TCGplayer repricing guide · undercutting strategy