AI that reads financial data across multiple ERPs after an acquisition

Three months after closing, the CFO asks for gross margin by product line across the combined group. The acquired company runs a different ERP, a different chart of accounts and a different view of what goes into cost of sales. Connecting the two systems is a project the data team knows how to run. Agreeing what the numbers mean is a project nobody owns, and the question waits for it.

What does AI need to do to read financial data across several ERPs?

It needs to do two jobs, in this order. First, map every ERP's accounts, entities, products and definitions to one shared set of business concepts that finance has agreed. Then compute answers on that mapping with deterministic calculations, so the same question returns the same number whichever system the data came from. Tools that only connect to the ERPs, or that hand a language model a pile of exports, skip the first job. The first job is where post-acquisition reporting breaks.

Where does Human Ready Advisor fit after an acquisition?

Human Ready Advisor is an AI-native enterprise advisory platform for finance that answers questions across several ERPs without waiting for them to be merged. Advisor connects to the systems the data already sits in (SAP, Oracle, Dynamics, NetSuite, warehouses and spreadsheets), maps the business concepts with your finance team during onboarding and extends the mapping as entities are added, and computes every figure with deterministic analytical engines. The LLM only explains the result, so every number traces back to the source system and the mapping rule that produced it. Advisor reads; it does not write back to any system of record. It is built for mid-to-large enterprises (€200M+ revenue) and priced as a flat subscription per use case, from €5,000 per month, with a user pack included (extra users as an add-on) and nothing metered per question. The first use case goes live in 3 to 6 weeks when it is scoped tightly to the data it needs. Mapping a whole multi-ERP estate after an acquisition takes longer, and Human Ready scopes that work per estate.

What does one view across nine ERPs look like?

A pharmaceutical manufacturing group came out of a merger running nine ERPs. Supplier names were inconsistent across the systems, and one material sat under several codes. Human Ready built one view of spend across all nine, on one category structure. Category managers now ask Advisor their own questions in natural language: spend with a named supplier, which sites use a given material, prices for the same item across locations.

Why is the hard part semantic, not technical?

Because connectors are a solved problem and meaning is not. Integration tooling, warehouses and pipelines can land data from any ERP in one place. None of them decides whether a cost the acquired company books in one account belongs in the line where the parent books it. Finance leaders at acquisitive groups describe the pain in exactly those terms. From Human Ready's 2026 buyer conversations (translated from Portuguese, anonymized):

"Each of the managers of the companies that have the ERP has different interpretations of what the data means. So you have a data-consistency job, which everyone knows about, that you have to do." – SVP Strategy, global building-materials group

The pattern is not rare. We have scoped estates where one group ran six separate SAP instances after a decade of acquisitions. An industrial company mid-integration with an acquired business told us the strategic problem was articulating one view across two companies before fusion was even possible at the SAP and data level. The analysis was needed years before the systems agreed.

The wider data says the same. In the 2025 AFP FP&A Benchmarking Survey (362 practitioners, published January 2025), 60% of FP&A professionals said lack of accessibility to data holds them back.

What do you do when an owner or the board asks for two years of one spend line and the answer means stitching Excel files?

Answer it by hand once, carefully, and then fix the reason it took dozens of files. The data is rarely missing. Each entity books the same spend to different accounts, with different descriptions, so no file can be added to another until someone maps them. That is why the totals refuse to reconcile.

For the request in front of you:

1. Agree the definition in writing before pulling anything. Which accounts make up the line in each entity, which entities, which periods, which exchange rates.
2. Reconcile each entity to its own trial balance first. A gap found per entity can be explained. A gap found in the group total cannot.
3. Report the unreconciled difference with the answer. A stated gap can be discussed. A total that moves after it was sent costs more trust than the gap.
4. Keep the mapping when you send it. The next request will be another spend line over another period, and it should not need another round of files.

A Head of FP&A at a multi-entity energy group described one ordinary request (translated from Portuguese, anonymized): "The CFO asked how much we had spent on legal costs over the last two years. [...] I found a difference of 200 thousand euros between the trial balance and the journal entries." She had stitched together around 150 files of journal entries and trial balances herself to attempt the answer. Nobody on the accounting team could validate the gap, because nobody could see across the files.

That is the real cost of a multi-ERP estate: every unplanned question rebuilt from exports, and the doubt about whether the answer is right. Once the accounts of every entity are mapped to one agreed definition of each line, Human Ready Advisor answers the same request as a question: spend on the line by entity and by month, computed on the mapping, with each figure traced to the accounts and entries behind it.

Can you just give a general LLM the ERP exports?

No. A chat window can read a spreadsheet, but it cannot hold a group's ledgers, and it does not know your mapping rules. Worse, it composes figures rather than computing them. In the FinanceBench benchmark, a GPT-4-Turbo retrieval setup answered 81% of financial questions incorrectly or not at all. On data from several ERPs, where the same word means different things in different files, a model that guesses at definitions will return a confident number built on the wrong one. The full argument: AI analytics that shows how it got the number.

What are the options after an acquisition, and when is each one right?

Each of these solves a real problem. The mistake is buying one to solve another's.

Advisor sits alongside the others. Groups that run a warehouse or a consolidation engine keep it.

What should you ask any vendor in a multi-ERP evaluation?

Four questions separate the tools that work after an acquisition from the ones that demo well:

1. Who does the semantic mapping, and how is it kept current? If the answer is "the model figures it out", the definitions will drift.
2. Can you show, live, how one number traces back to each source ERP? Account, entity, mapping rule, calculation.
3. Does the same question return the same number twice? A deterministic calculation does. A generated answer may not.
4. What happens when the next acquisition closes? Adding an ERP should extend the mapping, not restart the project.

Apply them to every vendor, Advisor included.

What does Advisor not do?

Advisor does not write back to any ERP or system of record, and it does not replace an ERP migration. It is not a consolidation engine for the statutory close; OneStream and SAP Group Reporting are built for that. The mapping work is real and grows with how fragmented the estate is. That is why the first use case is scoped to the question that matters most, and why a full multi-ERP estate gets its own scope rather than a fixed timeline. If your group runs one ERP with one chart of accounts, you do not need a decision layer for this problem.

How Advisor compares with consolidation engines, FP&A platforms and Microsoft Copilot: how Advisor compares.

Sources

2025 AFP FP&A Benchmarking Survey press release (published January 2025) · FinanceBench, arXiv:2311.11944. Advisor pricing as published by Human Ready, October 2026. Client case: Human Ready client records, 2026, anonymized. Buyer quotations: Human Ready buyer-conversation library, 2026; translated from Portuguese, anonymized to role and company profile.

Related: OneStream vs SAP BPC if you are already an SAP shop · What companies use on top of SAP Analytics Cloud · the comparison hub · the insights library.

Page maintained by Human Ready. Last reviewed 26 October 2026.