Deca Is Betting That Small Businesses Deserve Better Than QuickBooks and Spreadsheets

The Macro: The SMB Back Office Is Held Together With Duct Tape

I spend a lot of time talking to small business owners, and there is one complaint I hear more than any other: the back office is a mess. They’re running accounts payable in QuickBooks, tracking inventory in a spreadsheet, managing procurement through email chains, and chasing down invoices with a combination of Venmo requests and passive-aggressive follow-ups.

This isn’t a technology problem in the traditional sense. The tools exist. NetSuite can do all of this. SAP can do all of this. But those are enterprise products with enterprise pricing and enterprise implementation timelines. A 20-person manufacturing company doesn’t have $50,000 for a NetSuite deployment and six months to get it configured. So they stick with QuickBooks and spreadsheets and hire a part-time bookkeeper who spends 15 hours a week on data entry that software should be handling.

The ERP market for small businesses has been underserved for decades. QuickBooks is fundamentally an accounting tool that’s been stretched to do things it wasn’t designed for. Xero is cleaner but still limited. Sage Intacct is expensive. The gap between “basic accounting software” and “full ERP” has never been properly filled for companies with 5 to 200 employees.

What’s changed is AI. Not in the hype-cycle sense, but in a practical one. The back-office tasks that eat small business owners alive are exactly the kind of repetitive, pattern-matching, document-processing work that modern AI handles well. Invoice matching, purchase order processing, inventory reorder calculations, payment reconciliation. These are all tasks where a well-trained model can do in seconds what a human does in hours.

The Micro: One Platform to Replace the Patchwork

Deca is building an AI-native ERP for SMBs. The scope is ambitious: accounts payable, accounts receivable, inventory management, and procurement, all in one integrated platform. The AI isn’t a bolt-on feature. It’s the foundation. The system is designed from the ground up to automate the core workflows rather than just help you navigate menus faster.

The company is based in San Francisco and is part of a recent YC batch. The team is building what they describe as a modern version of the back-office stack, where AI handles the repetitive processing and humans handle the exceptions and decisions.

The product tackles the four pillars of SMB back-office pain. AP automation means invoices get processed, matched to purchase orders, and queued for payment without someone manually keying in line items. AR automation means invoices go out on time and follow-ups happen automatically. Inventory management means stock levels update in real time and reorder points trigger without someone checking a spreadsheet. Procurement means purchase orders flow through an approval process instead of getting lost in someone’s inbox.

I think the competitive positioning is smart. They’re not trying to compete with NetSuite for enterprise accounts. They’re not trying to be a better QuickBooks. They’re targeting the companies that have outgrown QuickBooks but can’t justify NetSuite. That’s a large market and it’s one where the incumbents aren’t paying close attention.

The risk I see is integration complexity. SMBs don’t operate in isolation. They need their ERP to talk to their bank, their payment processor, their e-commerce platform, their shipping provider. Building one great product is hard enough. Building one great product that connects cleanly to 30 other tools is significantly harder. Odoo has been trying to do this for years with an open-source approach and it’s still rough around the edges in several modules.

The Verdict

Deca is attacking a real problem in a large market with a modern technical approach. The SMB back-office gap is well-documented and poorly served. If the AI automation actually works, if it can process invoices accurately, manage inventory reliably, and handle procurement workflows without constant human babysitting, then this is a product that sells itself through word of mouth.

The challenge is trust. Small business owners are conservative about their financial systems. They need to trust that the AI won’t misclassify an expense, double-pay a vendor, or lose track of inventory. Building that trust takes time, and it requires the product to be boringly reliable before it can be impressively smart.

I want to see customer case studies in 60 days. Not vanity metrics, but specific numbers. How many hours per week does a 30-person company save? What’s the error rate on automated invoice processing compared to manual? If Deca can answer those questions with real data from real customers, they’ll have something powerful. If the answer is still “we’re building toward that,” the clock is ticking. The AI-native ERP space is going to get crowded fast, and being early only matters if you’re also good.