NATURAL LANGUAGE QUERIES
How to Query Acumatica Data in Natural Language Instead of Generic Inquiries, Reports, or Spreadsheets
Photo by Luke Southern on UnsplashThe old way to answer a finance question in Acumatica is to build a Generic Inquiry, run a report, or export to a spreadsheet. The new way is to ask. Natural language querying turns a three day request into a three message conversation, with the answer sourced back to the underlying records.
- Natural language querying lets anyone ask Acumatica a question in plain English and get a sourced answer without building a Generic Inquiry or export.
- Acumatica AcuChat, in early access in 2026 R1, answers questions in text, chart, or table form with drill through to records.
- A finance native layer adds cross object reasoning, calculated metrics, and the ability to act on the answer, not just display it.
- Every answer should cite its source records so finance can trust and verify it.
- ChatFin turns questions into sourced answers and, when needed, into approved actions inside Acumatica.
A surprising amount of finance time is spent turning a simple question into a query. Someone wants open payables by vendor, and a person builds a Generic Inquiry, exports to Excel, and formats a table. The answer was simple; the retrieval was the work.
Natural language querying removes the retrieval. You ask, the system finds the data, and it answers, showing its sources. This page covers what native Acumatica offers, where a finance native layer goes further, and how to keep the answers trustworthy.
The cost of the old way
Generic Inquiries, reports, and spreadsheets are powerful, but they put a builder between the question and the answer. That builder is a bottleneck. Questions wait for someone with the skill and the time, and by the time the export is formatted, the moment has often passed.
Photo by Priscilla Du Preez on UnsplashWhat Acumatica AcuChat gives you
Acumatica's own AcuChat, in early access with 2026 R1, lets users ask questions about ERP data directly in conversational English and receive answers as text, charts, or tables. Results are drillable, so you can click from an answer to the underlying records, and you can create dashboards from the conversation.
This is a real step forward for self service reporting inside the ERP. It answers questions well. The next step is reasoning across objects and acting on what you find.
Where a finance native layer goes further
Photo by Stefan Stefancik on UnsplashTrust: every answer needs a source
Finance cannot act on an answer it cannot verify. The rule for any natural language layer is that every figure traces back to the records that produced it. A good answer is not a paragraph of confident prose; it is a number with a link to the transactions behind it.
The standard to hold. If you cannot click from the answer to the source in Acumatica, do not post on it.
Who gets to ask
Natural language querying widens the audience. A controller, a budget owner, or a chief executive can ask directly instead of filing a request. That reduces the analyst's ticket queue and speeds decisions, as long as access respects the same field level security Acumatica already enforces.
From question to action
Photo by Kelly Sikkema on UnsplashThe most valuable questions imply an action. Which bills are due within the discount window? Which accounts look unreconciled? A finance native layer can carry the answer forward into a governed draft, so the insight does not die in a chat window.
How ChatFin answers Acumatica questions
ChatFin lets your team ask Acumatica finance questions in plain English, reasons across objects, returns calculated metrics with the source records attached, and, when the question calls for it, prepares an approved draft to act on. It complements AcuChat: use AcuChat for embedded, in screen questions, and ChatFin when the answer needs to become work.
The point of asking a finance question in plain English is not a nicer report. It is skipping the report and getting to the decision.
Frequently Asked Questions
Can I query Acumatica data in plain English?
Yes. Acumatica AcuChat, in early access with 2026 R1, answers questions about ERP data in conversational English as text, charts, or tables with drill through to records. For cross object reasoning, calculated finance metrics, and the ability to act on the answer, a finance native layer such as ChatFin adds capability on top.
Do I still need Generic Inquiries if I can ask in natural language?
Generic Inquiries remain useful as governed data sources and for structured feeds, and natural language layers often use them under the hood. What changes is that end users no longer need to build one for every question. They ask, and the system retrieves and answers, showing the source records.
How do I trust a natural language answer from finance data?
Insist that every answer cites the underlying records so you can verify it. A trustworthy layer returns a number with a link to the transactions behind it and respects Acumatica field level security. ChatFin attaches source records to answers and, for anything that posts, routes a draft through human approval.
Ask Acumatica, Get a Sourced Answer
ChatFin lets your whole team ask Acumatica finance questions in plain English, returns sourced answers with calculated metrics, and can carry the answer into an approved action.
Deploy on your existing Acumatica instance without replacing it, without database credentials, and without waiting for a development project. ChatFin reads through the same governed API surface Acumatica already exposes and writes back only drafts that a human approves.