AI Layer on Top of ERP: Inside Canals’ $35 Million Raise

Canals, a Miami-based AI software firm, raised $35 million in a Series A round led by Base10 Partners on May 28, 2026, betting that the future of business software is not a wholesale replacement of ERP, but a productivity layer that sits on top of it. More than 100 wholesale distributors already run on the platform. Canals had processed over 8 million sales orders and more than $5 billion in payables before taking a dollar of institutional capital. The new money is an accelerant on a thesis the company has been proving in production for years.

The round lands at a moment when the broader enterprise-AI conversation has tilted toward replacement. The argument now circulating in software is that foundation models will eat incumbent applications the way the internet ate retail. Canals is going the other way: it positions itself as an intelligence layer that reads messy inputs from emails, PDFs, spreadsheets, and voice, then writes structured data back into Epicor, Infor, SAP, and other systems of record. The bet, in plain language, is that ERP stays and AI rides on top.

How the $35 Million Round Lands

Base10 Partners led the Series A. The May 28 funding announcement and Base10 statement framed the deal as one of the largest AI raises to date in wholesale distribution. The company is profitable. It grew without institutional capital until this round. According to Canals’ company overview and team bios, the firm remains 90% employee-owned. The raise is meant to widen an Operating AI platform that already runs across sales, customer service, accounting, purchasing, and receiving at 100 distributors in the United States. Jason Kong, General Partner at Base10, put the due-diligence view on the record. “Canals is customer-obsessed and delivers outstanding ROI,” he said in the release. “Its technology and team are clearly the best in the market.”

The capital is earmarked for product development rather than market entry. Canals already counts DSG, The Kendall Group, and Locke Supply among its customers, alongside what the company describes as some of the largest distributors in the world. The next products in the pipeline target the friction between distributors and their suppliers: first-of-their-kind solutions for transactions that today still move by phone, email, and PDF. The 8 million orders and $5 billion in payables already processed give the new product work a production scale most AI startups in distribution do not have.

• $35 million Series A / May 28, 2026 / Base10 Partners led
• More than 100 wholesale distributors on the platform
• Over 8 million sales orders processed to date
• Over $5 billion in payables processed to date

Why an AI Layer Beats an AI Replacement

Michael Delgado, co-founder and chief executive of Canals, has a one-line description of what his company sells. The company, he said, is an AI layer. ERP vendors, he said, are systems of record. The distinction is the spine of the whole business. A layer reads documents in the formats customers already use, extracts the structured information those documents contain, and writes the data back into the ERP the customer has already paid to install. A replacement would require the customer to migrate. The layer approach accepts the install as a sunk cost and tries to make it more productive.

Most large distributors have lived with their ERP for a decade or longer. Migration projects are expensive, slow, and politically loaded inside the customer organization. A layer that bolts onto Epicor, Infor, or SAP is a different decision entirely.

Delgado also makes a labor argument for the layer model. The automation Canals delivers, he said, has not cost anyone a job. The quote, from a previously published interview: “To our knowledge, no one has ever lost their job because of the automation we’ve provided, but definitely folks have not had to be replaced because of the automation we’ve provided.” The framing matters: a layer that helps a sales rep turn quotes around faster reads as a productivity tool inside the existing organization, not a workforce reduction program. And if the layer itself fails, the customer falls back to the world they already know. Their ERP is still there. Their data is still there. They go back to manual entry until the vendor fixes the issue.

What Distributors Are Already Reporting

The customer metrics Canals published alongside the raise are reported by the company and by named customers, not independently audited. Read them as outcomes, not guarantees. The headline figures from the May 28 release: doubled quote conversion rates, 96% touchless invoice processing for early users of AI Accounts Receivable, and one customer whose win rate on automated quotes tripled against its pre-automation baseline. The doubling and the 96% figures come from Canals’ own press materials; the tripling was cited by Delgado in an earlier published interview.

Throughput before the round was the more striking number. Canals says it has processed 8 million sales orders and $5 billion in payables since launch, with sales growth that the company characterizes as surpassing best-in-class tech benchmarks. The growth came without venture capital. The May 28 round is the first institutional raise. The combination of profitability, bootstrapped scale, and now a Series A is unusual in a sector where most AI vendors are still chasing pilots.

On April 1, 2026, Canals broadened what the platform actually does. The April launch of Canals’ Operating AI suite added capabilities that take the layer beyond sales order entry into the back office. The new modules read supplier documents, match payments to invoices, answer customer inquiries from inside the ERP, and transcribe spoken orders into structured line items in real time. Early performance figures from the launch blog: up to 80% time savings for purchasing and receiving teams using AI PO-to-Receipt Tracking, and up to 99% match accuracy for AI Accounts Receivable.

One named customer closes the section. Paul Kennedy, CEO of DSG and Chair of the Board of Directors of the National Association of Electrical Distributors, made the comment in the May 28 funding release.

As DSG has expanded to serve customers across electrical, HVAC, plumbing, and waterworks, technology has always been a key enabler of our growth. Canals is now our gold standard for how technology can continue driving the business forward. It not only increases productivity so we can grow more efficiently, but also improves the customer experience and employee-owner experience.

The Speed Argument Behind the Layer

ERP vendors move slowly by design. Their system is the customer’s system of record, which means every patch carries the risk of breaking the books, the order pipeline, or the warehouse. The vendors absorb that risk by shipping conservatively. Canals makes the opposite trade: ship fast, fix fast, push new model capabilities the same day they ship.

Michael Delgado, CEO and Co-Founder of Canals, framed the contrast in his own words.

They release a bug patch in six months. We release a bug patch in 30 minutes. Our customers can take advantage of the latest and the greatest the minute something comes out so that they’re always getting better and better and faster.

The speed argument is what makes the layer model defensible against ERP vendors themselves. If Epicor, Infor, or SAP decides to build a comparable AI capability into the platform, they will move at their own cadence, measured in quarters, not minutes. Canals can ship the same capability faster, learn from customer data faster, and replace the feature before the customer’s next planning cycle. A layer that ships every day is harder to displace than a feature baked into a release roadmap.

The Risks of Betting on a Thin Layer

The wager has a structural vulnerability that the funding does not address. Canals sits on top of Epicor, Infor, and SAP, three vendors who control the customer relationship at the procurement table. If any of them ships a built-in AI layer that handles the same workflow, Canals loses its reason to exist inside that customer’s stack. The 30-minute patch advantage shrinks if Epicor’s native AI feature is bundled into a maintenance contract the customer is already paying for.

The customer base has its own concentration. The named reference customers are DSG, The Kendall Group, and Locke Supply, all of them electrical and plumbing distributors. A construction downturn would hit Canals twice, once through reduced order volume and once through reduced customer count. The vertical depth is the source of the company’s product accuracy, but it is also the source of its downside. 96% touchless invoice processing in one industry is not the same product as 96% touchless invoice processing across ten.

Size is the third risk. Canals is about 100 people. It grew without venture capital. The team is now accountable to a board that has not previously existed in the company’s history, and the capital brings expectations the bootstrapped years did not. If the layer thesis does not spread from wholesale distribution into adjacent verticals (HVAC services, industrial MRO, building products) the addressable market is bounded by the size of one industry. A $35 million Series A is meaningful inside distribution and small inside enterprise software.

The Pattern Beyond Wholesale Distribution

An American Supply Association report, cited in Delgado’s January 2026 Supply House Times column, found that nearly 75% of companies in the distribution channel are now experimenting with AI. The breakdown: 38% exploring use cases, 34% piloting tools, 19% implementing AI for operations like data forecasting and process automation. Canals is the product; the underlying pattern is the bet.

The wager is being made across the channel in the same shape. AI vendors are lining up next to installed ERP systems, reading the documents ERP cannot read, and writing structured data back into the fields ERP already exposes. Wholesale distribution represents $8.2 trillion in annual flows through the real economy, per the GlobeNewswire release. If the layer thesis holds, Canals processes a larger share of those flows, and the next $35 million round comes with a different multiple. If the ERP vendors close the layer into the platform, Canals becomes the company that proved the thesis and was acquired for the team. The 8 million orders and $5 billion in payables processed before the raise are the evidence either way.