Data foundations, reconciled before any model or system acts on them.

Each destination sets its own checks.

Readiness is a set of checks that pass or fail. We run them on your own records.

A migration ends when both systems agree.

Someone other than the person who ran the load confirms the match. Every migration is a script that runs first into a staging copy of the new system, is reconciled and runs again until the differences are zero. Production sees nothing before that.

• Record counts for customers, suppliers, items and open documents.
• The trial balance at the cutover date, account by account.
• Receivable and payable ageing, customer by customer and supplier by supplier.
• Stock quantities and valuation by warehouse, where you hold stock.
• A sample of historical documents opened in both systems and compared.

The scripts are idempotent: a re-run creates what is missing, reconciles what exists and leaves correct records alone. When a total does not reconcile, the script refuses to write. It rounds exactly as the target system does, because half-up and banker's rounding differ by a cent on some lines, and the cents add up. Our own books are held to the same rules.

On ERPNext, master records load through the Data Import tool or the REST API. Unpaid customer and supplier invoices go through the Opening Invoice Creation Tool, so due dates and ageing survive the move.

We migrate from anything that exports.

That covers QuickBooks, Zoho Books, Tally, Odoo, older ERPs and the spreadsheets that grew up beside them. Where an export falls short, we read the reports or the database and reconcile from those.

You decide how much history comes across. The usual choice is master records, open items and balances, with older transactions in a searchable read-only archive. Where reporting or audit needs several years of transactions in the new system, that load is scoped and priced on its own line.

Structured invoices start from clean master data.

Under the UAE e-invoicing mandate, 20 of the Ministry of Finance's 51 mandatory fields for a tax invoice come from company, customer and address records. Your Accredited Service Provider rejects what is wrong in them, so the cleanup comes first. The readiness checklist has the full count.

• TRNs. 15 digits, on every VAT-registered customer and supplier, held in one format.
• TINs. The first 10 digits of the TRN. The Ministry uses the TIN as the participant identifier, written 0235 followed by the TIN, and a business in scope with no tax registration obtains one through EmaraTax.
• Tax group members. Each member uses its own TIN, never the representative's.
• Addresses. Complete, and held in structured fields instead of free text.
• Tax treatment. Standard-rated, zero-rated, exempt, out of scope or reverse charge, set once on the item and the customer instead of typed onto each invoice.
• Duplicates. One customer, one record. Two records for the same buyer means two sets of invoices to reconcile.

A doubtful TRN is checked on the Federal Tax Authority's public TRN verification service before it reaches an invoice.

Documents ready for a model to find and read.

Gartner predicted in February 2025 that through 2026, organisations will abandon 60% of AI projects unsupported by AI-ready data. For a company whose knowledge lives in documents instead of a data platform, readiness means knowing what exists and where, in what format, how sensitive it is and who may see it.

• Corpus survey. An inventory of the drives, mailboxes and systems that hold the answers.
• Ingestion and OCR. Scanned and photographed documents pass through OCR, and quality is measured on samples.
• Permissions. Each document keeps who may see it, so a role that may not read a file never has it retrieved into an answer.
• Freshness. Superseded versions are marked or removed, so the assistant does not quote last year's price list.

You receive a data readiness read, which says whether your records can answer the questions you care about and what it would take to get them there.

Data warehouses sit outside this service.

Data readiness here means migration, cleanup, reconciliation and ingestion into the systems we build and run. Data warehouses, platforms such as Snowflake or Databricks and large business intelligence programmes are out of scope, and if you need one, that comes up on the first call.

Everything the migration produced stays with you.

• A mapping document.

  Every source field, its target and the rule that joins them.

• An exception list.

  Each record that could not be migrated or cleaned automatically, with the reason.

• A reconciliation report.

  Counts, trial balance, ageing and stock, old against new, signed off by your finance lead.

• The scripts.

  They pass to you with the final payment and can be run again.