Integration
Dynamics 365 Business Central runs finance, inventory and fulfilment once a sales order exists. LevelOps PDF to Order handles what comes before: emailed PDF and Excel purchase orders become validated Shopify draft orders, approved by a person, then carried into BC by the Shopify Connector you already use.
Dynamics 365 Business Central is the mid market ERP for companies that already live in Microsoft. Finance, inventory, warehousing and fulfilment sit in BC, dimensions carry the analytics, and the rest of the estate is right there: Outlook, Teams, Excel, Power BI, Power Automate and Dataverse. The pull of that stack is real, and it shapes how every problem gets solved.
On the commerce side, Business Central ships a first party Shopify Connector. It syncs customers, items and orders, and imports Shopify orders as BC sales orders. So the standard architecture is settled: Shopify is where the order is created, BC is where it is fulfilled and invoiced, and the connector moves it across.
Which leaves one question nobody in that architecture answers. What creates the order when a wholesale buyer emails a PDF? Today the answer is a person, working through an attachment in Outlook and typing it into Shopify or straight into BC. LevelOps PDF to Order replaces that keyboard step with AI extraction, catalogue matching and a human approval, and then hands the order to the connector you already trust.
These are the complaints specific to a BC shop, not generic data entry pain.
The PO arrives in a shared mailbox. Somebody triages it, opens the attachment, and retypes it. That workflow is invisible to BC, so it never appears in any process map or improvement plan.
BC lets you record customer item numbers, and it is a good feature. Keeping it current across hundreds of accounts is a data chore nobody owns, so reps translate the buyer's codes from memory instead.
Your analytics depend on dimensions being right. When an order is keyed under pressure, the dimension defaults get accepted without thought, and your Power BI reports inherit the guess.
Teams build flows that move attachments around and create tasks. The flow still ends at a human who has to decide which item a line refers to, because a purchase order is not a structured form.
A mistyped quantity that reaches a posted shipment or invoice becomes a credit memo, a return order and a reconciliation. Catching it at intake is an order of magnitude cheaper.
Extending BC means AL development, a sandbox, testing and a deployment window. Meanwhile the orders keep arriving, and the interim solution is the same two people and the same inbox.
Nothing changes on the BC side. The order simply arrives in Shopify already correct, which is the one thing the Shopify Connector cannot do for you.
Forward it to your dedicated LevelOps address from Outlook, set a mailbox rule, or upload the file from the Shopify admin.
Items, quantities, units, requested dates and the PO number are read off the document, then each line is matched to your Shopify catalogue, which the Shopify Connector keeps aligned with your BC item master.
The result is a standard Shopify draft order with the source document attached. Your rep checks the flags, corrects anything wrong, and approves. Nothing exists until then.
Once approved it is an ordinary Shopify order, so the first party connector imports it as a sales order with your existing customer and item mapping, and BC takes over from there.
This is the honest shape of the integration today. LevelOps writes to Shopify natively, and order output for Business Central is in progress rather than released. If BC has to be the destination on day one with no Shopify step, wait for it.
The field level division of labour between the document, Shopify and Business Central.
Extraction happens before anything is created.
A normal draft order built from your Shopify data, not from the PDF.
Whatever the Shopify Connector already maps. These orders are not a special case.
The boundary is part of the design.
In a Business Central shop, BC is the item master and the Shopify Connector publishes a subset of it to Shopify with prices, inventory and variants. PDF to Order matches against that Shopify catalogue on purpose, so every matched line points at something the connector can carry back into a BC sales order without a manual fix.
The practical consequence is worth stating plainly. If your buyers order items that exist in BC but were never synced to Shopify, those lines are flagged for a person instead of matched. That is correct behaviour, and it usually reveals a sync scope decision made years ago that nobody has revisited. Widening the item sync, or agreeing those items are genuinely off catalogue, is short work.
Customer item numbers are where BC users feel this most. Item References are the right long term hygiene, but they only help for codes somebody has already recorded. Matching handles the buyer's own part numbers, abbreviations and free text descriptions against your catalogue with a confidence score, so a code nobody has ever seen still resolves on the first order, and anything genuinely ambiguous is flagged rather than defaulted. Your team's corrections feed back into the matching, so accuracy climbs on your data rather than on someone else's benchmark.
The app stands on its own. Business Central is also one of the platforms our implementation team delivers, so the deeper work is available from the same partner.
We sample the purchase orders your team processed last month, group them by account and format, and estimate real coverage. You get a number based on your documents, not a case study.
We confirm which items, customers and prices are in sync between BC and Shopify, and where the connector's scope leaves gaps that would show up as flagged lines.
Start with the accounts that send the messiest documents. Every draft is reviewed by your team, corrections train the matching, and you measure against your own baseline.
If the connector mapping, dimensions or warehouse setup need work, that is a Business Central implementation engagement. See the ERP and MRP implementation page for how those are scoped and run.
Business Central is one of the five platforms our implementation team delivers on, with fixed fee scope and defined milestones. Detail at /en/erp-implementation.
Everything here is about the document that arrives before a sales order exists. The Shopify Connector is not a competitor, it is the step after.
| Manual entry in BC | Power Automate and AI Builder | EDI programme | LevelOps PDF to Order | |
|---|---|---|---|---|
| Reads any PO layout without configuration | A person does | Per form training | Per partner mapping | Yes |
| Matches loose descriptions to your items | From memory | Rarely | Mapped up front | Yes |
| Applies the account's contracted price | If the rep remembers | No | Yes | Yes |
| Human approval before the order exists | Yes | Depends on the flow | Usually automatic | Yes |
| Handles a layout change next month | A person adapts | Retrain the model | Change request | No template to break |
| Needs BC development | No | Often, plus Dataverse work | Usually | No |
| Time to first live order | Immediate, forever | Weeks of build | Months per partner | Days |
| Where the order is created | Shopify or BC by hand | Wherever the flow posts | Straight into BC | Shopify draft, then the connector |
Rated 5.0 on the Shopify App Store. Same pattern, different stacks.
Before LevelOps, B2B orders were entered manually or with Excel macros. Manual processing made the workload very unpredictable, especially on days with large orders that took enormous amounts of time.
After using LevelOps for three months, the percentage of email PDF orders requiring manual entry has dropped to under 10%, down from 60% prior to adopting the tool.
Draft orders get entered by magic. Even when the PO contains a vague description and no SKU, this app can still match it. The AI is that good.
Such a timesaver and super easy to implement. We process purchase orders via EDI, PDF, Excel or email and the entire process has been seamless.
Not today. Order output for Business Central is in progress, not released, and we would rather you plan around what exists. What ships now is native Shopify draft order creation. Your BC sales order is then created by the Shopify Connector for Business Central, which is a first party Microsoft feature, or by whatever middleware you already run.
It does the heavy lifting on the ERP side. Business Central ships a first party Shopify Connector that syncs customers, items and orders, so once a draft order is approved in Shopify it becomes an ordinary Shopify order and the connector imports it as a BC sales order. LevelOps only fills the step before that, which is turning a document into a correct Shopify order in the first place.
BC Item References are genuinely useful and worth maintaining, but PDF to Order does not depend on them. Matching resolves the buyer's own part numbers and loose descriptions against your Shopify catalogue directly, which means a new customer code works on the first order rather than after somebody adds a reference record.
Power Automate is excellent at routing and orchestration, and AI Builder can read a consistent form. Purchase orders are the opposite of a consistent form: every account has a different layout, and layouts change without warning. The hard part is not the flow, it is matching a loose line to the right item and price with a confidence score, and refusing to guess. That is the part we built.
Yes. The PO number is extracted from the document and written to the Shopify order, and the Shopify Connector maps that into the External Document No. field on the BC sales order, which is where your team expects to find it. The source file stays attached to the Shopify draft for audit.
Exactly as they do for your other Shopify orders. Dimensions are resolved in Business Central from the customer, item and location by your existing setup and default dimension rules. If a dimension depends on something written on the purchase order, tell us during the audit and we confirm the field is captured before you commit.
No. The item master, customer master, posting groups and all financial postings stay in BC and we never touch them. LevelOps writes to Shopify only, which means there is nothing to certify, no AL extension to install and no BC permission set to change.
It works if your Shopify to BC mapping already routes orders to the right company, typically one Shopify store or sales channel per BC company. Intake behaves the same in all of them, because the routing decision happens after the Shopify order exists rather than during extraction.
If Shopify is part of your order flow and something already carries Shopify orders into NAV, yes, because we only write to Shopify. If nothing bridges Shopify and NAV today, that bridge is the project, and it belongs in an implementation engagement rather than an app subscription.
Send us three real purchase orders from your hardest accounts. We run them live and show the extraction, the matched lines and the flagged exceptions before you commit. Every store also gets your first 5 orders free, with no expiry.
Business Central is one of the platforms our implementation team delivers.
Read, match, price, approve, in detail.
Why creating a Shopify draft first is safer than posting into an ERP.
Where EDI earns its cost, and where email intake is cheaper.
The same Shopify first pattern for merchants on Oracle NetSuite.
B2B and DTC on one store, wired to your back office.
Bring three real POs from the accounts that give your order desk the most trouble. We run them live, show the draft orders and the flagged lines, and tell you honestly how much of your volume this covers today.
your first 5 orders free, free allowance of 5 orders, bilingual support in English and French.
