Company
LevelOps started as an operations and systems consultancy for manufacturers, distributors and wholesalers. The same gap showed up in every engagement, so we built the product that closes it. This page is the practice, the method and what we believe. The founder's career that produced it is on Our story.
LevelOps is a small Montreal team of operations and systems people. Before there was a product, there was a practice: implementing ERP and MRP systems, and building on Shopify, for manufacturers, distributors and wholesalers.
That work is unusually revealing. An implementation is not a slide deck, it is a month inside somebody's actual operation, watching the order flow from the moment a customer sends something to the moment cash arrives. You see which steps are documented and which ones live in one person's head. You learn that the process map on the wall and the process on the floor are different documents.
We did that repeatedly, across industries that had almost nothing in common: supplements, apparel, robotics, food and beverage, building materials, sewing patterns. Different products, different systems, different countries. And the same failure, in almost every single one.
The systems were fine. Shopify handled the store. The ERP handled production and finance. The B2B portal handled the accounts that were willing to use it. Every piece of software did what it said on the box.
But the orders themselves still arrived as PDFs and spreadsheets in an inbox, and a person retyped them. Line by line, between phone calls, from thirty different customer formats, into a system that had no idea any of it had already been typed once by the buyer who sent it.
What made it worse was how invisible it was. Nobody's dashboard showed it. It was not a line item in a budget. It was absorbed by customer service as part of the job, and it was the single largest recoverable cost in the flow. Every implementation plan we wrote, and every one we reviewed from someone else, quietly assumed that step would be handled by somebody else. So nobody handled it.
PDF to Order came out of that pattern rather than out of a product strategy. There was no market sizing exercise. There was a list of clients with the same problem and a reasonable guess that if we solved it once, properly, it would work for all of them.
What it does is narrow on purpose. It reads whatever a customer sends, as a PDF, an Excel file, a CSV, a Word document, a scan or plain text in an email body. It matches each line against the seller's Shopify catalogue, including when the buyer uses their own part numbers or a loose description with a typo in it. It applies the price list that account was actually promised, never the number printed on the buyer's document. Then it creates a Shopify draft order and stops, so a person can review it with the source file attached.
The consulting work still funds and informs the product. That is why the product starts at the inbox rather than at a dashboard, and why the first thing we offer a merchant is an audit rather than a trial.
The same four steps whether we are implementing an ERP or onboarding a merchant onto PDF to Order. Order matters more than speed.
Map how orders arrive, who touches each one, how long it takes and what it costs. This step exists so that nobody, including us, gets to guess where the bottleneck is.
Design the flow first, then the systems around it. A tool installed on top of an undecided process just makes the confusion faster.
Build it, roll it out, and stay through the part where reality disagrees with the plan. One merchant described four weeks of structured onboarding and joint iteration, which is what this step actually looks like.
Keep the catalogue and customer data clean, then put automation where it earns its keep. AI on top of bad master data produces confident nonsense.
The audit is free and comes with no obligation, because a wrong diagnosis is the most expensive thing in this business.
Six positions we have taken. Each one closes off options we could otherwise sell, which is rather the point.
Every order lands as a draft with the source file attached. The failure mode of a confident wrong order is expensive, and the trust does not come back. A review step is not a limitation we plan to remove.
Asking a wholesaler's buyers to log into a portal loses orders from exactly the accounts least willing to change. We read what they already send instead.
The inbox, not the dashboard. Reporting on a broken process is a slower way of confirming what the person doing the typing already told you.
Merchants told us 60% of orders needing manual entry fell below 10%, and that a 100 item order now takes under four minutes. Those are their numbers, on their files, which is the only kind worth quoting.
We use AI where it removes repetitive error prone work and nowhere else. Novelty is not a business case, and one merchant calling us their first easy step into AI is worth more than any demo.
If intake is not your bottleneck, we will say so during the audit. Being able to help two ways, as consultants or with a product, is what makes that honesty affordable.
Two things that feed each other. We implement ERP and MRP systems and build on Shopify for manufacturers, distributors and wholesalers, and we build PDF to Order, an app that turns inbound purchase orders into Shopify draft orders. The consulting work is how we learned where order flow breaks. The product is the part we could turn into software.
Montreal, Quebec. We work in English and French with wholesalers and distributors across Canada, the United States and Europe, and support is bilingual on every plan.
Because one step kept falling outside the scope of every implementation. Consulting fixes a process one client at a time, and there are problems where the honest answer is that the client needs software that does not exist yet. Order intake was one of those. Building it once, properly, was better than writing the same workaround into a thirtieth statement of work.
Because that is where the work actually is. In engagement after engagement the systems were fine and the orders were still being retyped by a person from a PDF. Every implementation plan quietly assumed somebody else would solve that step, so nobody did.
A confident wrong order is more expensive than a slow one. It ships the wrong product, costs roughly $250 to unwind, and spends credibility with a customer that does not come back. So every order lands as a draft with the source file attached, and a person has the last word.
Yes, and deliberately. Implementation work keeps us inside real operations, which is where we find out what the product is missing. It also means we will tell a merchant when automation is not their bottleneck, because we have a way to help them either way. It is also the older half of the business: the practice came first, and the product's requirements came out of it.
An operational efficiency audit, then workflow and system design, then custom development and implementation, then ongoing data management support and AI. Proven, transparent and stress free is the promise, and the audit comes first because guessing is expensive.
We do not ask a wholesaler's customers to change how they order, we do not build automation that acts without a review step, and we do not promise a number we have not measured on real orders. Those three rules cost us some deals and save us all the interesting ones.
On Our story, told period by period from before 2006 to the rename. This page is the practice that career produced. That one is the rooms it was learned in.
Book a free order flow audit. Twenty minutes on how orders reach you, who touches each one and where they stall. You keep the map and the numbers whether or not you ever install anything.
The same three faces as Our team. Office photographs of Square-Victoria are not in the repo yet. Drop the files and they go in a frame under this block.
CEO
Maxime Lavoie, president and founder of LevelOps, has spent the last 15 years helping manufacturing and retail companies improve their operations and technologies. Maxime was Vice President of Operations at Goodfood and a supply chain consultant at Deloitte
LinkedIn profile for Maxime Lavoie
CTO
Thierry Ferland specializes in automation and system integration. Former CTO of La Récolte, a French farm to table retailer, he develops technological solutions to improve the efficiency of manufacturing and retail businesses.
LinkedIn profile for Thierry Ferland
AI Solutions Engineer
Mathieu Magri develops and fine-tunes the AI models behind product matching and document processing. That is the part of PDF to Order that has to read a purchase order it has never seen before and still find the right variant in your catalogue.
LinkedIn profile for Mathieu MagriThe career, period by period, from the maple farm to Ax.c.
The people, the roles and the experience behind the work.
Associations, backers, service firms and the marketplaces that list us.
What changed for merchants, with their numbers rather than ours.
Read, match, price, approve, with a person keeping the last word.
Twenty minutes on your order inbox, with an honest assessment.
We will map the flow, put numbers on the manual steps, and tell you plainly whether automation is your bottleneck. If it is not, we will say so.
Montreal based. English and French. No pitch, no commitment.
In numbers
