AI ecommerce analytics answers a seller's plain-language questions about profit, margin, ads, inventory, fees, and cash from that seller's own reconciled data, across the channels the data covers. The twelve questions below are the ones sellers running Amazon plus two or three other marketplaces ask before they trust it. Answers refer to Crunch, the analytics AI inside ConnectBooks, where the answer depends on the product; otherwise they apply to the category.
It works across the channels whose data has been reconciled into one ledger on one cost basis. ConnectBooks syncs Amazon, Shopify, Walmart, eBay, TikTok Shop, and Temu, and the integrations page lists each with its QuickBooks and Xero pairings. The Crunch page names Amazon, Shopify, Walmart, TikTok Shop, and eBay as the channels it reads. If a channel's settlements post to your books as one lump per deposit with no fee split and no SKU detail, no AI can answer a SKU question about it, whatever the marketing says.
Yes, and "clean" has a specific meaning here: each marketplace settlement matched to its bank deposit and broken into order, fee, refund, and adjustment lines; landed cost per unit entered with freight; ad invoices booked by channel; returns tied to the original order. An AI on books that are 90 percent there will answer 90 percent of questions correctly and the other 10 percent with the same confidence. How to prepare your books for an AI CFO covers the checklist. Skip it and the first useful thing the AI does is expose what is missing.
Their numbers stay outside the analysis. If you sell on a marketplace that is not synced, the AI's "total profit" is total profit on the synced channels, and any question about channel mix is answered on a partial catalog. The honest handling is to label the answer as partial. Ask the tool what it does when a channel is missing; a flag is the right response, a silent total is the wrong one. Meanwhile, manual journal entries for the unsynced channel keep your P&L complete, even though the AI cannot see inside them at SKU level.
It sees ad spend that has been booked into the ledger. Amazon Ads invoices, Walmart Connect charges, and TikTok Shop ad and affiliate deductions each arrive in a different shape and land in the books through the channel sync or the accounting connection. ConnectBooks includes a PPC ROI view built on booked ad cost and attributed sales. What no AI on ledger data sees is the campaign console itself: keyword bids, placements, search terms. For "did the ad spend on this SKU earn its keep," the ledger is enough. For "which keyword," open the console.
It answers questions the dashboard has no tile for. A dashboard shows the metrics its designer chose: margin by SKU, ad spend by month, inventory by warehouse. It cannot answer "which products were profitable last quarter but lose money this quarter, and why." An AI on the same data can, because it runs the comparison, isolates the SKUs, and explains the cause. Sellers keep the dashboard for the daily glance and use the AI for the question that would otherwise take an afternoon in a spreadsheet.
Only as well as the returns were booked. A refund on Amazon posts with its referral fee reversal; a refund on eBay may carry a fee credit; a Shopify return costs a label; a TikTok Shop return on an affiliate order raises a commission question. If each return posts against its original order and SKU on its channel, the AI can report return rate by SKU by channel and the cost of each return. If returns sit in one account with no order link, it can report a total and nothing more. Ask it "which SKU has the highest return rate on Walmart" and see whether it answers or guesses.
Yes, provided the SKU map is in place and COGS is shared. A product listed under an ASIN on Amazon, an item ID on Walmart, and a product ID on TikTok Shop is one SKU in the books only if someone mapped those identifiers. Once mapped, and with one FIFO cost applied to every channel's units, the AI can show contribution per channel for the same product and name why they differ: referral versus commission, FBA versus label, ad model, return rate. Without the map, the same product is three products, and the comparison is impossible for any tool.
As current as the last sync and the last settlement. Marketplace settlements arrive on each marketplace's cadence, so "profit this week" is an estimate until the settlement that covers it lands, at which point it becomes a fact. Ad invoices post on the platform's billing cycle. Landed cost is as current as the last PO entered with freight. An AI answering "what happened yesterday" is working from orders and estimated fees; an AI answering "what happened last month" is working from settled numbers. Know which one you are asking.
It can answer cash questions that the ledger supports: what the next payouts look like based on settled orders, how much cash sits in inventory by SKU, and, for Q4 decisions, what a SKU's incremental cash contribution is if held, discounted, liquidated, or removed. Crunch treats already-paid COGS as sunk cost in that view and weighs storage fees, aged-inventory surcharges, expected sell-through, and expected cash recovery. It cannot see the bank balance in real time, the loan covenant, or the supplier who wants a deposit next Tuesday. Those are yours to bring to the answer.
The Crunch page calls it an AI CFO for ecommerce sellers, and the questions it answers (profit, margin, ads, inventory, fees, cash) are the ones a fractional CFO would field. It is not accountable for the close, does not sign off on anything, cannot call the supplier, and does not know the deal you struck with your 3PL last week. It is an analyst that works from reconciled books and answers in seconds. The judgment stays with the person who reads the answer.
Pricing depends on monthly order volume and the number of marketplaces connected, and the tiers are Gold, Diamond, and Platinum. The pricing page has the current breakdown and which capabilities sit in each tier. There is a 30-day free trial and no free plan.
Start with one question you have wanted answered for months and could not get from a dashboard. "Why is profit down this quarter on a flat top line" is the usual one. Then check whether your books can support it: settlements reconciled at the transaction level, landed cost entered, SKUs mapped across channels. If they can, connect the channels and ask. If they cannot, the first project is reconciliation, and the AI is the reward for finishing it. Book a demo to see the question run on sample data before you commit your own.
An AI on reconciled data reads what settled. It does not know the supplier raised prices before the new PO hits the books, it cannot judge whether a low-margin SKU is worth keeping for the customers it brings in, and it does not act in any seller portal. It answers. You decide.
Running an e-commerce business comes with plenty of challenges, but ConnectBooks is here to make your life easier. With real-time insights, seamless integrations, and detailed tracking of your profitability and inventory, you can stay ahead of the game. Whether you’re selling on Amazon, Shopify, Walmart, TikTok or eBay, ConnectBooks helps you manage your finances with 100% accuracy and confidence, so you can focus on growing your business.
Ready to level up? Start making smarter, data-driven decisions every step of the way. Try ConnectBooks Free Today or Schedule a Demo
