Customer service
Customer emails answered with real order details
Questions about deliveries, returns and products get a drafted reply that already includes the tracking status, the order history and your policy.
Starts when
A customer email or chat message arrives.
What happens
- Identify the customer, the order and what they are asking
- Look up tracking, stock and return eligibility
- Draft a reply in your tone that resolves the question in one message
- Escalate complaints, damage claims and anything involving a refund
In more detail
An automation like this reads each customer email or help-desk ticket, works out which order it is about, and looks up the real order, tracking status, return window and stock in the store and shipping systems. It then drafts a reply in the shop's tone that answers the question with those facts, such as the carrier status and expected date. Complaints, damage claims, refunds and late orders go to a person with a summary. A staff member reads each draft and sends it; in the design Fenrir Labs could offer, nothing sends on its own. It needs a written returns and shipping policy, orders with tracking numbers, and read access to the store platform and help desk. Late orders matter legally: the FTC Mail Order Rule requires offering a delay-or-cancel choice, so those get a special draft.
You stay in control
You choose which replies send automatically and which wait for a person.
Connects to
- Help desk
- Store platform
- Shipping carriers
Questions we would ask you
- Where do customer messages arrive today: Gmail, Outlook, Gorgias, Zendesk, Help Scout, Shopify Inbox, marketplace inboxes?
- Roughly what share are order-status questions versus returns and product questions?
- Which store platform and shipping tool hold the orders and tracking numbers?
- What is your written returns and shipping policy, and how often do you make exceptions?
- Which replies must always be read by a person before sending?
Discuss this further with the Live Agent
Marketing & reviews
Product listings from supplier sheets
A supplier's spreadsheet and photos become complete listings: titles, descriptions, attributes and categories, formatted for each sales channel.
Starts when
A new supplier file or product batch is added.
What happens
- Read the specifications and match them to your attribute fields
- Write titles and descriptions in your store's style
- Check for missing dimensions, materials and compliance details
- Format the listings for your store and each marketplace
In more detail
An automation like this takes a supplier's spreadsheet or line sheet and the product photos, maps the supplier's columns to the store's own fields, and drafts titles, descriptions, tags and categories in the store's style. It checks each product for missing facts such as dimensions, materials, care instructions or required warnings, and flags them instead of guessing. The output is a set of unpublished draft listings, or an import file, plus a report of what is new, what changed and what is missing. A merchandiser reviews the batch and publishes it. It needs a style guide or a few example listings, consistent SKU rules and a list of required attributes per product type. The first version can work purely from files without connecting to the store.
You stay in control
A merchandiser reviews the batch and publishes it.
Connects to
- Supplier files
- Store platform
- Marketplaces
- Photo library
Questions we would ask you
- How many new products a month, and from how many suppliers?
- What formats do suppliers send: CSV, XLSX, PDF line sheets, portal downloads?
- Which channels do listings go to, and do they differ per channel?
- Do you have a style guide or a set of listings you consider ideal?
- Which product types carry compliance details (fiber content, warnings, age grading)?
Discuss this further with the Live Agent
Operations & inventory
Reorder suggestions that account for lead times
Each week you get a purchase plan built from how fast each item is selling, how long the supplier takes and what is already on order.
Starts when
A weekly schedule, or stock falling below a set level.
What happens
- Calculate the selling rate for each item, adjusted for season and promotions
- Compare stock on hand and on order with supplier lead times
- Suggest order quantities that respect minimums and case sizes
- Draft a purchase order per supplier
In more detail
An automation like this builds a weekly purchase plan. It reads sales and stock from the store platform and point of sale, adds what is already on order, and uses each supplier's real delivery time, minimum order and case size to suggest how much to order of each item. The arithmetic is ordinary code; the AI only writes a short explanation per supplier. The plan is grouped into draft purchase orders that the buyer edits and places; nothing is ordered automatically. It needs inventory tracked per item, a list of suppliers with lead times and case sizes, and purchase orders recorded somewhere. Ready-made tools such as Prediko and Inventory Planner already do this well for Shopify stores, so a custom design mainly suits shops with data in several systems. Shopify's Stocky app was retired in August 2026, so many shops are choosing a replacement now.
You stay in control
The buyer adjusts quantities and places each order.
Connects to
- Point of sale
- Store platform
- Inventory
- Supplier email
Questions we would ask you
- Which systems hold sales and stock: Shopify, Square, Lightspeed, a spreadsheet?
- Did you use Stocky, and what replaced it after August 2026?
- How many active SKUs and suppliers?
- Do you know each supplier's real lead time, minimum order and case size?
- How do you record purchase orders and receipts today?
Discuss this further with the Live Agent
Customer service
Returns processed from request to refund
A return request is checked against policy, a label is issued, and when the parcel arrives the refund and restock are prepared for approval.
Starts when
A customer requests a return.
What happens
- Check the order date, item and reason against your return policy
- Issue a label and instructions, or explain why the item is not eligible
- Record the condition when the item is received
- Prepare the refund or exchange and update inventory
In more detail
An automation like this handles a return from request to refund. It reads the customer's request, finds the order, and checks the purchase date, item and reason against the shop's written return policy. Eligible returns get a drafted message with a return label; ineligible ones get a polite explanation citing the policy. When the parcel arrives, staff record its condition, and the system prepares the refund, store credit or exchange and the stock update. A staff member approves every refund above the amount the owner sets and every exception, and the money only moves after that click. It needs a clear written policy, a shipping label account and a receiving routine. For Shopify stores, ready-made apps such as AfterShip Returns and Loop already cover most of this, so Fenrir Labs could design the email-reading and drafting part around them instead.
You stay in control
Staff approve refunds above the amount you set and every policy exception.
Connects to
- Store platform
- Shipping carriers
- Inventory
- Payments
Questions we would ask you
- How many returns a month, and what share are exchanges?
- What is your written policy, and how often do staff make exceptions?
- Which platform handles refunds, and do you sell on marketplaces with their own return flows?
- Who pays return shipping, and which label tool do you use?
- Who inspects returned items, and how is condition recorded?
Discuss this further with the Live Agent
Marketing & reviews
Reviews monitored and replies drafted
New reviews across every site are collected in one place with a drafted reply, and recurring complaints are reported so they can be fixed at the source.
Starts when
A new review is posted.
What happens
- Collect reviews from your store, marketplaces and map listings
- Draft a reply that addresses what the reviewer actually said
- Connect negative reviews to the order so the problem can be put right
- Summarize recurring themes by product each month
In more detail
An automation like this gathers new reviews from the store's review app, Google and other sites into one list and drafts a reply to each that responds to what the reviewer actually said. Negative reviews are linked to the order where possible so staff can put the problem right privately, and reviews mentioning damage or late delivery can open a support ticket. A person approves and posts every public reply. Each month it summarises recurring complaints by product. It never writes or buys reviews, and it does not filter who is asked for a review, in line with the FTC's 2024 rule on consumer reviews. Google requires the owner's express consent and an approved API project before replies are posted through its API. It needs owner access to each review site, a reply style guide and someone checking the queue every day or two.
You stay in control
A person approves each public reply.
Connects to
- Review sites
- Store platform
- Help desk
Questions we would ask you
- Where do your reviews appear: your store app, Google, Amazon, Etsy, Trustpilot, Facebook?
- Which review app do you use on the store, and on which plan?
- How many new reviews a month, and how many get a reply today?
- Who is allowed to reply in public, and in whose name?
- What can staff offer an unhappy customer without asking you?
Discuss this further with the Live Agent
Finance & billing
Supplier invoices checked against agreed prices
Each supplier invoice is compared with the purchase order, the goods received and your agreed price list, and discrepancies are raised with the supplier.
Starts when
A supplier invoice arrives.
What happens
- Read the invoice lines and match them to the order and receipt
- Compare unit prices with the current price agreement
- Calculate the difference for short shipments and price errors
- Draft a dispute or credit request with the evidence attached
In more detail
An automation like this reads each supplier invoice that arrives by email and checks every line against three things: the purchase order, what was actually received, and the price agreed with that supplier. Lines that match are marked clean. Where a unit price is higher than agreed, units were billed but not received, or a discount or free-freight term is missing, it lists the difference and drafts a credit request with the evidence. The buyer approves and sends each dispute; clean invoices go to the bookkeeper. Nothing is paid or entered into the accounts automatically. It needs purchase orders, a record of deliveries received and current price lists per supplier. Because supplier prices are confidential and the work is mainly reading and comparing, it suits a model running on the shop's own computer.
You stay in control
The buyer approves each dispute; clean invoices move to payment.
Connects to
- Email
- Inventory
- Accounting
- Price agreements
Questions we would ask you
- How many supplier invoices a month, and from how many suppliers?
- Do suppliers send PDFs, paper, or portal downloads?
- Do you raise purchase orders for every order, and where are they kept?
- How do you record what was received?
- Where are agreed prices kept, and how often do they change?
Discuss this further with the Live Agent
Sales & leads
Wholesale inquiries qualified and quoted
Trade inquiries get a fast reply with the right price tier, minimums and lead times, and are followed up until the buyer orders or declines.
Starts when
A wholesale inquiry arrives by form or email.
What happens
- Identify the business and check it against your trade account criteria
- Build a quote from the correct price tier and current stock
- Send the quote with terms and a trade application if needed
- Follow up on a schedule and stop when the buyer replies
In more detail
An automation like this reads each wholesale inquiry from the website form or email, pulls out the business name, location, type of shop and products of interest, and checks it against the seller's written trade criteria. It then builds a quote from the right price tier, minimum order, case sizes, current stock and lead times, and drafts a reply with the quote, terms and a trade application if needed. Prices come from the seller's own price list by ordinary code, not from the AI. A sales lead approves larger quotes and any custom pricing, and a person sends every message. Follow-up drafts are scheduled and stop when the buyer replies; follow-ups count as commercial email under CAN-SPAM. It needs written trade criteria, a price list per tier and current stock levels. Because wholesale prices are confidential, it suits a model on the seller's own computer.
You stay in control
A sales lead approves quotes above a set value and all custom pricing.
Connects to
- Website form
- Email
- Price tiers
- CRM
Questions we would ask you
- How do wholesale inquiries arrive today, and how many a month?
- What makes a shop a good or bad fit for your brand?
- How many price tiers do you have, and where are they kept?
- What are your minimum order, case sizes and payment terms?
- Do you sell through Faire or Shopify B2B, and do you prefer one channel?
Discuss this further with the Live Agent
Reporting
The morning sales briefing
Before opening, you get one page: yesterday's sales by channel, what sold out, which ads paid for themselves and what looks unusual.
Starts when
Every morning at a time you choose.
What happens
- Pull sales, refunds and stock levels from each channel
- Compare with the same day last week and last year
- Point out stock-outs, sudden movers and advertising that is underperforming
- Write the briefing in plain language and send it to your phone
In more detail
An automation like this sends the owner one page each morning before opening. It reads yesterday's sales, refunds and stock from the point of sale, the web store and any marketplaces, and spend and results from the ad accounts. Ordinary code calculates the totals, the comparisons with last week and last year, items that sold out or soon will, and which ads paid for themselves. The AI then writes it up in plain language, and every number is checked against the calculated figures before the page goes out. It only reads data and changes nothing, so there is nothing to approve. It needs read access to each system, consistent channel tagging and working ad conversion tracking. Shops that sell only on Shopify may find Shopify's own reports and Sidekick assistant enough; a custom briefing helps most when sales are spread across several systems.
You stay in control
Nothing to approve; it is a read-only report built from your own data.
Connects to
- Point of sale
- Store platform
- Ad accounts
- Email
Questions we would ask you
- Which channels do you sell through, and which numbers do you check first each morning?
- Which POS and store platform, and do you sell on Amazon, Etsy or Faire?
- Which ad accounts do you run, and is conversion tracking set up?
- Do you want comparisons with last week, last year, or a target?
- Email, text or a phone notification, and at what time?
Discuss this further with the Live Agent
