ERP&BusinessManagementSystems

Modules I build

### Inventory and warehouse Multi-warehouse and multi-branch stock, transfers with approval, batch and expiry tracking, serial numbers where they matter, stock-take with variance reporting, reorder points, and landed-cost handling so your margin figures survive contact with shipping and customs.

### Purchasing and suppliers Purchase requisitions and orders, supplier records and price lists, goods-received notes matched against orders, supplier ageing, and three-way matching between order, receipt and invoice.

### Sales and customers Quotations, orders, delivery notes, invoicing, customer ageing, credit limits, price lists per customer type, and a CRM pipeline that a salesperson will actually update because it takes ten seconds.

### Accounting A real double-entry general ledger — not a report generated from a sales table. Chart of accounts, journals, accounts receivable and payable, bank reconciliation, fixed assets and depreciation, cost centres per branch, and trial balance, P&L and balance sheet that tie out.

Built around Kurdistan's actual tax position: 15% flat corporate income tax on net income, a 30 June filing deadline, no VAT and no sales tax. Your books come out in a shape your accountant can file from, rather than one they have to rebuild.

### HR and payroll Employee records and contracts, leave and attendance, shift scheduling, and payroll that handles the local reality: the 1,000,000 IQD monthly tax-free allowance, income tax withholding above it, and 12% employer / 5% employee social security. Payslips in Kurdish, Arabic or English.

### Multi-currency The same discipline as the POS work: prices held in IQD or USD, an owner-controlled house rate with a change log, and the rate snapshotted onto every transaction so historical reports stay stable. Reports available in either currency, with the basis stated on the report.

### Four working languages — including both Kurdish dialects Every module runs in Bahdini, Sorani, Arabic and English, and each user picks their own at login. A warehouse clerk in Duhok works in Bahdini while the accountant in Hawler works in Sorani and the owner reads reports in English — same database, same records, one system.

Bahdini and Sorani are treated as two separate languages, not one checkbox. Most vendors here ship "Kurdish" meaning Sorani only, which is no use to a company operating in Duhok, Zakho or Akre. Invoices, payslips, purchase orders and reports all render correctly in whichever is chosen, with right-to-left layout built in rather than patched on.

### Reporting and dashboards Branch-by-branch P&L, margin by product and by customer, stock valuation, cash position, ageing reports, and an owner dashboard that answers "how are we doing" in one screen instead of one week.

### AI, where it actually pays for itself I build AI automation and AI agents as well as the systems themselves — which means the AI sits on top of your real ledger rather than guessing from an export.

The use cases that consistently earn their cost in an ERP:

• Invoice and document extraction into accounts payable — supplier invoices, delivery notes and receipts read automatically and matched three ways against order and receipt. This is the most reliable return in the entire category, and almost nobody here is doing it yet
• Demand forecasting and reorder planning — seasonality, Ramadan and Newroz peaks, and supplier lead times factored into what you should be holding
• Ask-your-data reporting — "what was our margin on this supplier's products across all branches this quarter" answered in plain Kurdish, Arabic or English instead of built by hand
• Exception detection — duplicate payments, prices that drifted from the price list, a customer quietly running past their credit limit, stock adjustments that need explaining
• Customer and supplier messaging — quotes, order confirmations and payment reminders drafted in the right language and tone, sent on WhatsApp, logged against the record

And the honest counterpart, which is also in my AI consulting work: a good deal of what gets sold as "AI for ERP" is a report with a chatbot in front of it. I will tell you which of the above is worth your money for your business and which is not — including when the answer is none of them yet.

Sold as modules, not as a monolith

You do not buy an ERP. You buy the two modules that are on fire, then add the next one when it earns its place. Every module is built on the same database and the same permission model, so adding the third one is a week of work rather than a new project.

A typical sequence:

1. Inventory + purchasing — stops the stock bleeding, usually pays for itself first 2. Sales + invoicing — one price list, one document trail 3. Accounting — the ledger the first two feed 4. HR + payroll — once the headcount makes it worth automating

Where this fits, and where it doesn't

Good fit: distributors and wholesalers, multi-branch retail, food and beverage groups, manufacturing with a simple bill of materials, construction and contracting with per-project costing, medical suppliers, companies running 10–250 staff.

Bad fit — I will tell you: if you are a single shop with one till, you want POS, not ERP. If your process is genuinely standard and you have the budget, an established product like Odoo may serve you better than anything bespoke, and I will say so — including which local partner to talk to. Custom is worth paying for when your workflow is your advantage, not when it is just undocumented.

Getting off what you have now

Nobody starts from zero. The migration work is part of the job: extracting from Excel, from an old Access or desktop system, or from a product whose vendor has disappeared — cleaning the master data, mapping the chart of accounts, and running the old and new system in parallel until the numbers agree for a full month before anyone switches off.

How pricing works

There is no price list, and you should be suspicious of any ERP vendor who gives you one. The same "ERP" can mean two modules for a 12-person distributor or seven modules and a two-year data migration for a 200-person group.

What I will tell you openly is what actually drives the number:

The main drivers

• How many modules — this is the biggest single factor, and it is why I sell them one at a time
• How many branches, warehouses and users
• Migration — how much data has to come across, and how clean it is. This is routinely underestimated and is often the largest line item
• Integrations — payment wallets, existing systems, hardware, anything that has to talk to something else
• How unusual your workflow is — standard processes are fast; a process that is genuinely your competitive advantage takes longer to model properly
• Multi-currency — full dual-currency with reconciliation is real engineering, not a settings toggle

What keeps it down

• Starting with two modules instead of seven
• Clean, exportable existing data
• One person on your side empowered to answer questions quickly — this affects timeline more than anything in the code

How you find out your number: a free scoping call, then a written proposal with fixed scope and a fixed price per phase. You see the breakdown before you commit, not after. If you decide to stop after phase one, you own everything built in phase one.

I will also tell you when the answer is *don't buy this from me*. If your process is genuinely standard and you have the budget, an established product may serve you better than anything bespoke — and I will say which one and who implements it locally.

Book a free 15-minute call →