Data to AI™ / ChatGPT/Exact Online
Accounts payable and receivable teams ask ChatGPT what is overdue, what is due out and where spend exceeds budget, and approve every Exact Online lookup it makes.
Better answers, fewer tokens.
ChatGPT
Which customers have the most money stuck in the oldest receivables age group?
Van Dijk Bouw has the most in the oldest group, €14,200 of its €31,500 open total. Across these three accounts, €24,950 sits in Age Group 4.
| Account Name | Total Amount | Age Group 1 Amount | Age Group 4 Amount |
|---|---|---|---|
| Van Dijk Bouw Financial - Aging Receivables List | €31,500 | €9,800 | €14,200 |
| Kessler Logistik Financial - Aging Receivables List | €18,750 | €6,300 | €6,450 |
| Maas Retail Financial - Aging Receivables List | €12,900 | €8,600 | €4,300 |
Read from the current aging list, which reloads all open items on every run. Bucket meaning comes from the Age Group Description fields.
Exact Online is cloud business software for accounting and financial management. Dataddo extracts its receivables aging, payables, sales invoices, budgets and profit and loss data.
Show Financial - Payables list entries with a Due Date in the next 7 days, grouped by Account Name, with Amount and Approval Status for each.
Before releasing payments, the AP lead checks which supplier invoices fall due and which still wait for approval, confirming each tool call so the source of every line is clear.
Run data_status first, then give me Overdue Receivable Invoice Amount and Overdue Payable Invoice Amount per Currency Code from the Outstanding Invoices Overview.
The finance manager opens the morning stand-up with overdue totals on both sides of the balance, and the freshness check means nobody quotes numbers from a stale run.
Compare Budgets AmountDC by GLAccount Description with the Amount in Financial - Reporting balance for this Reporting Year, and list the five largest overruns.
A budget owner finds the G/L accounts that ran ahead of the budget scenario, then asks follow-up questions in the same chat about the accounts with the biggest gap.
Data to AI™ is priced per seat, like the AI assistants your team already uses, with ~25% off on annual billing. Need tailored contracts or volume pricing? Talk to us about Enterprise.
Serve governed business data to Claude, ChatGPT and other AI tools over MCP.
$10.49/month
per seat
3 seats minimum
Flexibly deploy Dataddo in any cloud or hybrid environment.
Custom
We have a payment model that works for you
| Without Dataddo |
With Dataddo
|
Outcome for you | |
|---|---|---|---|
| Preparing the payment run | The AP lead downloads the payables list and the outstanding invoices overview, then uploads both files to ChatGPT before every payment run. | A developer-mode app reads Financial - Payables list and the overview from Dataddo whenever the lead asks, with no files involved. | No uploads before payment runs |
| Knowing what was read | Once a file is uploaded, nobody can tell later which export or which tab a figure in the reply came from. | ChatGPT asks you to confirm each tool call, so you see which Exact Online entity it describes or queries before anything runs. | Every lookup visible |
| Snapshot lists | A payables export pasted into the chat looks current even when it was taken before last week's payments were booked. | Lists such as the Payables list reload all current data on each scheduled run, and data_status tells ChatGPT when that run happened. | Dated answers |
| Bank account fields | Exports of Payments carry Bank Account Number and Account Bank Account Number columns straight into a chat file. | Untick those fields when you create the source, so they never reach the flow and ChatGPT never receives the account numbers. | Sensitive fields stay out |
| Token consumption | ChatGPT spends tokens exploring raw Exact Online columns, sampling rows and retrying until it understands the data. | Field definitions and relations are supplied up front, so the model skips the exploration and goes straight to the answer. | Fewer tokens per answer |
| Answer quality | Prone to wrong joins, invented fields and numbers that are hard to check. | Answers grounded in defined fields. In the Dataddo benchmark, 78.5% of questions were answered correctly, against 65.5% with plain CSV files. | Better answers |
The Exact Online datasets finance and RevOps teams use most, with their real field names. ChatGPT queries them by name through the Dataddo semantic layer.
| Dataset | Metrics and attributes |
Ask ChatGPT
|
|---|---|---|
| Financial - Aging Receivables List |
|
How is the Total Amount of open receivables split by Currency Code? |
| Financial - Payables list |
|
Which suppliers by Account Name do we owe the largest Amount? |
| Sales - Invoices |
|
Which customers by Invoice To Name have the largest invoices due this month? |
| Financial - Outstanding Invoices Overview |
|
How has Outstanding Receivable Invoice Count changed over the last few months? |
| Financial - Profit and Loss |
|
What is our margin this period when comparing Revenue Current Period with Costs Current Period? |
| Budgets |
|
How does budgeted AmountDC differ between each Budget Scenario for the same year? |
We asked 43 pre-registered questions of three real datasets - Google Search Console, HubSpot deals and Google Ads - and delivered the same rows to the same model in four ways. Only the data layer changed.
of questions answered correctly - the best of the four delivery methods tested
fewer confidently wrong answers - 14.0% of runs against 24.0% and 26.3%
more correct answers on messy CRM data - better on 8 questions, worse on none
Same model, prompts and rows in every condition; 759 graded runs, ground truth frozen before any run. On Google Search Console alone, documented CSV files did as well as Data to AI™ - the advantage comes from messier data such as HubSpot deals.
Read the benchmark →Create an Exact Online source in Dataddo with datasets such as Financial - Outstanding Invoices Overview, Financial - Payables list and Budgets, and attach it to a ChatGPT destination or AI Model. In ChatGPT on the web, turn on developer mode, create an app for the remote MCP server https://headless.dataddo.com/mcp-data with OAuth and sign in.
Plans with developer mode: Pro, Plus, Business, Enterprise and Education, on the web. On Business and Enterprise a workspace admin may first need to allow developer mode before you create the Dataddo app.
ChatGPT asks you to confirm each tool call before it runs. You see whether it is listing flows, describing the Payables list or running a query, and you can decline any call you do not expect.
Yes. In the app settings you can turn individual Dataddo tools on or off, for example switching off refresh while keeping query, data_status and known_gaps available for everyday questions.
Only if you extract them. Payments contains fields such as Bank Account Number. Untick them when you create the source and they never reach the flow, so ChatGPT never receives those values. The rest of the Payments data stays available.
Lists without a date range, such as Financial - Aging Receivables List, reload all current data on each run, so values change only after the next scheduled extraction. Ask ChatGPT to run data_status to see when that was.
Developer mode, which the Dataddo app relies on, is available on ChatGPT on the web only. Set up the app and ask your Exact Online questions in a web chat.
