Illustration titled What large companies run on. Four tiles: one record for each thing, all the history, problems get fixed, and a team that changes it. A bar at the bottom reads: Now built around a small business.
Figure 1. The four things a large company's system gives it.

What large companies get from one model of the business

Many people are becoming familiar with ontologies in business thanks to the success of companies like Palantir. The idea underneath is simpler than the word. An ontology is one shared description of what the business deals with and how those things relate: customers, products, orders, purchase orders, invoices and suppliers, and the links between them. What an ontology is, explained for business owners covers it in more detail. When a company's systems are built on that description, it gets four things.

  • One record for each thing. Sales, the warehouse, purchasing and finance read the same customer, the same order and the same stock count. When a number is questioned, everyone is looking at the same number.
  • All the history. Years of orders stay in the system staff use, so the company can see which customers are buying less, what sells in which season and what really happened on a disputed order.
  • Problems get listed and fixed. A record that doesn't add up lands on someone's list and is corrected from a source document, instead of surfacing later on an invoice.
  • A team that changes it. A new pricing rule, product line or report becomes a change to the system, not a side spreadsheet.

Here is an illustrative example of the difference. A customer disputes an invoice from two years ago. At a 15-person wholesaler running a CRM, an order system, accounting software and a few spreadsheets, the invoice is in the accounting software, the order behind it is in an old system only one person still logs into, and the customer's substitution request is in a sales rep's inbox. Answering takes two people most of a morning. At a company with one model, the invoice, the order, the purchase order that filled it and the note are linked records, and one person answers in minutes.

Why small businesses have gone without it

Small businesses have wanted this for a long time. The cost has been in four places.

  • Building it. A system set up around one business's work has meant a custom software project, billed by the hour and lasting months.
  • Moving the history. Bringing years of orders, customers and invoices across with their links intact is slow, careful work, so it is usually the first thing cut from a quote.
  • Checking the history. Proving the moved data is right, field by field, takes longer still, so most migrations are checked by looking at a handful of records.
  • Changing it afterward. Every change after launch becomes a new quote, so changes stop and workarounds start.

The wholesaler in our case study shows where that leads. It ran on Zoho for 15 years, and an outside consultant added a customization every time the business needed something Zoho did not do. Each change solved the problem in front of it. After 15 years only the consultant fully understood how they fit together, and much of the data also lived in spreadsheets, exports, backups and folders outside Zoho.

What changed: an ontology, agents and automation

Three changes have brought the cost down without lowering the standard.

Describing the business is now the fast part. Building the model first, and fitting the software to it, used to be a consulting engagement. For the wholesaler, the ontology took a couple of days. The method is not new or untested. Intelligence Factory's published case study on integrating dozens of EHR systems describes data from nearly three dozen EHR and EMR systems mapped to one internal ontology, with the mappings for each source kept separate from the rules that use the data. Don't copy your old database: design the data model first explains why the model comes before any data moves.

Agents do the entry and the checking. We build and change CommerceServer with Buffaly, our own software engineering system. Buffaly agents read documents into the model, check them against its rules and write the records, and anything that fails a check goes to a person. When a file arrives in a layout the agents have not seen, Buffaly writes a new reader for it and uses it in the same session. We consider this key to how Buffaly works. In Buffaly Builds Its Own Tools, an analysis of 1.2 million messages and over 380,000 tool calls from a real Buffaly instance found that 70% of newly created tools were used in the same session they were written. Agents change records only through checked actions, as can you trust an AI agent with orders, prices and invoices? explains.

Moving and checking history is automated. An import that runs in minutes can be run again after every correction, and every field can be compared with a backup of the old data. Why your data import should be rerunnable explains the approach.

Illustration titled What changed. Three blocks joined by arrows: Ontology, describe the business; Agents, do the entry and checks; Automation, import in minutes. The last arrow points to a building labeled Small business. A bar at the bottom reads: The cost came down. The standard did not.
Figure 2. Three changes brought the cost down for a small business.

What it looked like for a flooring wholesaler

The wholesaler left 15 years of Zoho customizations in one month.

  • One model. The ontology described its customers, products, inventory pieces, vendors, orders, purchase orders and invoices, and the CommerceServer schema was changed to fit it.
  • All the history, checked. An automated pipeline imported 15 years of records, hundreds of thousands of them, in minutes. Every field brought across was compared with a backup of the old data: 293,986 fields, with zero differences after 43 exceptions were traced back to source documents.
  • What the business needs. Every requested feature and report was built, plus a B2B portal where customers check inventory and submit purchase orders. The business now runs its own custom workflow, custom business rules and custom reporting, all supported in CommerceServer. It never had that in 15 years on Zoho.
  • No double entry. During a one-month parallel run, staff kept working in Zoho and emailed the PDF Zoho produced, and Buffaly agents entered the work in CommerceServer.

The business is now off Zoho, no longer needs the consultant and no longer pays Zoho license fees. A sister company followed right away. What was learned moving the first company became the tools for the second: the import method and the checks run before and after it. How a flooring wholesaler left 15 years of Zoho patches in one month tells the full story, and the second migration was the easy one covers what came next.

Where to start: six questions

Answer these about your own business.

  1. If a customer disputes an order from three years ago, can one person answer it without asking anyone else?
  2. Does each customer exist exactly once across everything you use?
  3. Can you see your full sales history, or only the years since your last software change?
  4. When data is wrong, is there a list of what is wrong and who is fixing it, or do you find out from customers?
  5. When you want the system to work differently, does it change within days, or do you start a spreadsheet?
  6. If your longest-serving employee left tomorrow, would the business still know what it knows?

Several "no" answers mean the business runs on effort and memory. Before talking to any vendor, list every place your data lives and the business need behind each customization, then ask how your history will be moved and checked and who will change the system after launch. The business data migration checklist sets out the checks to ask for.

Tell us how your business runs today

CommerceServer is built, run and continually improved by one accountable team, with the improvement capacity agreed for each engagement. If your business runs on several apps, a consultant and the memory of long-serving staff, tell us what happens today and what needs to change. We will tell you what one model of your business would involve. Tell us about your current system, or read how full service works.