It's Tuesday. A capsule drops on Thursday, 400 SKUs still need checking, the Korean copy isn't back yet, and your biggest stockist is waiting on a linesheet. Any one of these is manageable. All of them at once, drop after drop, is a lot — and it lands on the same few people every time.

In our last guide, we said AI works best when you start from a real business problem. This guide picks up from there. We've taken a practical look at the everyday jobs AI agents can already take off your team's plate, including a super-simple breakdown of some playbooks you can start testing today.

Merchandising lives in the details: every attribute filled in, every colourway represented, every size chart consistent, every market priced. With these playbooks, your AI agent checks those details the way your most careful merchandiser would, and tells you exactly what's missing or wrong.

Launch readiness check

Try asking. "Is the Aurora coat ready to launch on our UK site on Thursday?"

What it knows. Every colourway needs its own photos. Every active size needs a barcode. And the product has to be sellable in that market, not just visible.

Does it change anything? Reports back only. You get one verdict: launch-ready, or what's left to do.

Size chart hygiene

Try asking. "Do we have duplicate size charts, and are our labels consistent?"

What it knows. Duplicates are matched on the actual size run, not the chart name. Before proposing a merge, it checks that a 38 means the same size on both charts, because merging an EU 38 onto a UK 38 quietly sends shoppers the wrong garment.

Does it change anything? Suggests a fix, but nothing changes unless you approve it. Each proposed merge or rename shows how many products it will affect.

Bulk attribute fill

Try asking. "Set fit to Regular on everything in the Denim folder that doesn't have a fit assigned yet."

What it knows. It only sets Regular if that's a fit value that already exists in your setup, so you don't end up with a near-duplicate like "Regular fit" next to it. And any denim that already has a fit is left as it is.

Does it change anything? Changes after you approve. It previews the batch, applies only what you approved, and rolls the whole batch back if anything fails.

There’s plenty more merchandising playbooks available through Centra MCP: Product sellability audit · Catalog completeness audit · New products review · Attribute filter search · Bundle integrity audit · Dead stock retirement · Slow-mover & markdown radar

Every new market multiplies the copy, translations and feed data that have to be right. With these playbooks, your AI agent keeps them in step, market by market, and shows you where the gaps are.

Product catalog translation

Try asking. "Localise the new SS27 drop for our Japanese site."

What it knows. It translates the copy shoppers actually see, not just the internal product record, and keeps your brand voice and glossary terms consistent. Where it isn't sure, it flags the text for your review instead of guessing, and it only suggests new product URLs if you ask it to.

Does it change anything? Changes after you approve. It shows you every translation before writing it, then checks that each one landed.

Missing translations report

Try asking. "What's still not translated into Japanese across the catalog?"

What it knows. It checks the text shoppers see, like product descriptions and category names, in every language your site is live in, so you can prioritise what matters most.

Does it change anything? Reports only. If you want, it can draft the missing translations as a next step, which you approve separately.

Product feed exclusion audit

Try asking. "Which products are falling out of our Google Shopping feed for the UK, and why?"

What it knows. It separates what gets a product dropped from the feed, like a missing title, price, image or identifier, from what only weakens the listing, like a missing description.

Does it change anything? Suggests a fix, but nothing changes unless you approve it.

There’s plenty more ecommerce & localisation playbooks available through Centra MCP: Category translation · Filter & attribute translation · Image alt-text translation · Copy product setup between stores

Prices and promotions run across markets, currencies and campaigns at once, and that's where small mistakes hide. With these playbooks, your AI agent checks the prices shoppers actually see, and makes campaign changes in bulk when you need them.

Price consistency audit

Try asking. "Which products are priced in EUR but missing in GBP or SEK?"

What it knows. It checks every active pricelist and treats retail and wholesale separately, so a product priced for your stockists but not for your own site still shows up as a gap. It also flags anything priced at 0.00, since that looks priced but sells for nothing.

Does it change anything? Suggests a fix, but nothing changes unless you approve it. The gaps come sorted by severity, with anything priced at nothing first.

Omnibus lowest-price check

Try asking. "Are any of our 'lowest price in 30 days' figures obviously wrong?"

What it knows. A price set to zero, even briefly, can be recorded as the lowest price and shown to shoppers as free. It flags those, and lowest prices far below today's price, before they mislead shoppers or an Omnibus report.

Does it change anything? Reports only. These records can't be corrected directly, so it tells you what to escalate. A clean result means no obvious errors, not a compliance sign-off.

Bulk campaign editor

Try asking. "Extend the mid-season sale by a week and raise the 30% tier to 40% on outerwear."

What it knows. It changes only what you asked for and leaves every other campaign setting as it was. It works on one campaign at a time, so a change never spills into another.

Does it change anything? Changes after you approve. It previews every change and writes nothing until you say yes.

There’s plenty more pricing & promotions playbooks available through Centra MCP: Campaign performance report · Campaign price impact report · Voucher expiry cleanup

Wholesale runs on seasons, stockists and their accounts. With these playbooks, your AI agent keeps the details behind every account right, from what a buyer can see in the showroom to getting a new stockist ready to order.

Wholesale sell-through

Try asking. "What's the sell-through on AW26 by market, and for our top five accounts?"

What it knows. Sell-through can be measured in more than one way, so it always names the formula it used. That way, nobody argues about the number.

Does it change anything? Reports back only.

Wholesale orderability check

Try asking. "Why can't Selfridges see the Aurora coat in the showroom?"

What it knows. It checks, in order, everything that decides whether a buyer can order a product: the account's market and pricelist, the delivery window, the price and the stock. Then it tells you the first one that's blocking.

Does it change anything? Suggests a fix, but nothing changes unless you approve it. You get the exact fix for the blocker it found.

B2B account onboarding

Try asking. "Set up our new stockist in Copenhagen on the EUR wholesale pricelist with 30-day payment terms, and invite their buyer to the showroom."

What it knows. A new account needs more than a name before it can place an order: a market, a pricelist, payment and shipping terms, and a buyer with a real billing address. It sets all of it up together, so the stockist can order in your showroom from day one.

Does it change anything? Changes after you approve. It previews the whole setup first, then creates the account and sends the buyer their showroom welcome email.

There’s plenty more wholesale playbooks available through Centra MCP: Wholesale season setup · Buyer order status report · Wholesale account health scorecard · Order discrepancy audit · Sales rep performance

Stock, orders and suppliers move every hour, and problems rarely announce themselves. With these playbooks, your AI agent finds what's short, stuck or about to run out, and helps you move stock to where it's needed.

Low-stock reorder triage

Try asking. "What do we need to reorder before it sells out, and what's already on the way?"

What it knows. It only counts stock that can actually ship, not stock sitting in your stores or already promised to orders. And it checks open purchase orders, so you don't reorder what's already coming.

Does it change anything? Suggests a fix, but nothing changes unless you approve it. It proposes purchase orders, but never places them.

Backorder triage

Try asking. "What's on backorder right now, and what will the incoming deliveries fill?"

What it knows. An order can look fine on paper while some of its items are still waiting for stock. So it checks every item on every order, not just the order's status, and sorts what's short into two groups: what incoming deliveries will cover, and what's stuck with nothing on the way.

Does it change anything? Reports only.

Redistribute stock between orders

Try asking. "Take 6 units of the black straight trousers from the pending Zalando order to complete the Selfridges order."

What it knows. It works out exactly how many units are short, names the order that would give them up, and won't touch orders that are locked or already being processed.

Does it change anything? Changes only after you approve. It previews the move, then re-checks both orders to confirm it worked.

There’s plenty more operations & supply playbooks available through Centra MCP: Allocation diagnostics · Sold-out diagnostic · Stale order triage · Order anomaly watch · PO receiving status · Store replenishment · Bulk order fulfilment

Like any new colleague, start your agent on one job. Pick a recurring task your team would happily hand over, find the playbook for it, and connect your AI tool to Centra with read-only access in a test environment. Run it a few times, see how the agent works, and widen its access when you're comfortable. Once your team trusts a few playbooks, bigger jobs follow, like getting a whole capsule live in a new market.

Centra MCP is live today. If you're already on Centra, you can set it up yourself with the install guide, or ask your Customer Success rep to help. If you're not, book a demo and we'll show you what your team could hand over first.

Every Centra MCP playbook, grouped by team. You can ask your agent for one by name, or simply describe the job.

Merchandising: Attribute filter search · Bulk attribute fill · Bundle integrity audit · Catalog completeness audit · CSV / spreadsheet import · Dead stock retirement · Launch readiness check · New products review · Product media upload · Product sellability audit · Size chart hygiene · Size-level orderability report · Slow-mover & markdown radar · Status disagreement audit · Sales channel launch checklist

Ecommerce & localisation: Category translation · Copy product setup between stores · Filter & attribute translation · Image alt-text translation · Missing translations report · Product catalog translation · Product feed exclusion audit

Pricing & promotions: Bulk campaign editor · Campaign performance report · Campaign price impact report · Omnibus lowest-price check · Price consistency audit · Voucher expiry cleanup

Wholesale: B2B account onboarding · Buyer order status report · NuOrder linesheet export · NuOrder order import · Order discrepancy audit · Sales rep performance · Wholesale account health scorecard · Wholesale account pricing change · Wholesale orderability check · Wholesale season setup · Wholesale selection status by SKU · Wholesale sell-through

Operations & supply: Allocation diagnostics · Backorder triage · Bulk order fulfilment · Customs readiness check · Ingrid booking gap scan · Inventory movement report · Low-stock reorder triage · Order anomaly watch · PO over-delivery check · PO receiving status · Redistribute stock between orders · Shipment tracking audit · Sold-out diagnostic · Stale order triage · Stale purchase order check · Stock anomaly scan · Store replenishment ·  Allocation request performance

Customer service & CRM: Compensation credit · Customer attribute search · Customer metrics · GDPR anonymisation sweep · Lost shipment resolution · Restock & back-in-stock signal · Return processing (RMA) · Serial-returner watchlist · Subscription health report · Voucher unlock preview · Wishlist review

Finance: Gift card overview · Gift-code liability tracking · Import tariff setup · Landed cost analysis · Order tax report · Payment reconciliation report · Refund & return reconciliation · Wholesale credit control · Order payment diagnosis

Reporting & support: Any-cut metric export · Custom report builder · Integration health snapshot · Support case handoff