MCP Data-Quality Report for Your AI Client

Analyse your data with your AI client · Data-quality report

Your AI client knows the data is dirty before it analyses it

Empty and blank values, inconsistent formats, duplicate keys, outliers and suspicious amounts, checked by your Network Agent over the whole result in one pass.

Business Core and up Runs on your Network Agent Alleen-lezen, altijd

Loop je ergens vast? Vraag het ons gerust. Open Ondersteuning Ga naar Query Streams (menu rechtsboven → Ondersteuning) en stuur ons een bericht. Echte mensen lezen het, en "Ik weet niet wat er op dit scherm wordt gevraagd" is een volkomen terechte vraag.

What your Network Agent does for you

Before your AI client builds an answer on a result, it can check that the result is trustworthy. Your Network Agent profiles the result and runs the quality checks in the same single pass inside your network, then gives the client a report of findings ranked by severity, not data.

Rekening	What it finds
Empty values	Columns that are empty, or mostly empty.
Blanks	Text filled in with nothing but spaces.
Format consistency	Dates, emails and codes that do not follow the format most of the column uses, with examples.
Duplicate keys	Rows that share what should be a unique key: the table’s primary key on a plain read of one table, or the columns you name.
Outliers	Values far outside the usual range of a column, with the bounds used.
Leading digits	Amount columns whose first digits do not follow the pattern real-world amounts follow (Benford’s law), a common sign of made-up or mis-keyed figures.

Situatie	Before	With a quality report
“Sum revenue by region.”	A confident total, with 12% of rows missing a region.	Your AI client says so first and asks how to treat them.
A migrated customer table	Duplicates found months later.	Duplicate customer numbers named on day one.
Hand-keyed expense amounts	Nothing looks wrong.	The leading-digit check flags the column for review.

The plan it needs

Plan	This capability	Opmerkingen
Gratis	Not on this plan	Works as it always has, with standard results.
Personal Core, Personal Plus	Not on this plan	Context management and Result analysis are included; Deeper analysis is not.
Business Core, Business Plus, Business Max	Inbegrepen	Part of Deeper analysis.
Enterprise Core, Plus, Max	Inbegrepen	Part of Deeper analysis.

Why it is on paid plans: these are analyses a chat client cannot do by reading rows. Your Network Agent runs them over the result it is holding, inside your network, and gives your AI client the groundwork to reason about: counts, scores, tests and examples.

The Free plan works as it always has. Paid plans add the Analytic toolset. Free is a plan you keep: MCP, the REST API, federated queries and alerts, with the same results as before. Plan names and prices are on the prijspagina.

How to ask

You do not call tools yourself. Ask in plain words and your AI client picks the tool. Three asks that bring this capability in:

Claude

ChatGPT

Cursor

Claude Code+10 moreAlleen-lezen

Check the quality of the customer table before we use it.

qs_run_query · mode=quality

Are there duplicate customer numbers?

qs_run_query · mode=quality · key_columns

Do the expense amounts look genuine?

qs_run_query · mode=quality · amount_columns

Groundwork done on your Network Agent, inside your network ingelogd op het tabblad MCP-activiteit

For people who read tool lists: mode: quality op qs_run_query en qs_run_saved_query, met key_columns, amount_columns en max_rows. See it on the Tools tab.

What comes back

Claude

ChatGPT

Cursor

Claude Code+10 moreAlleen-lezen

Before we analyse inventory costs, is the data any good?

qs_run_query · mode=quality · max_rows=200000

Checked 200,000 rows across the table’s columns. The issues, most serious first:

ernst	issue	kolommen
kritisch	empty in every row	several legacy columns
warn	many values missing	batch_ref
warn	leading digits off the expected pattern	unit_cost, material_cost
warn	mixed date formats	received_on
informatie	no key columns named, duplicate check skipped

The two cost columns are the ones to look at before totalling anything. Want examples of the date formats?

Groundwork done on your Network Agent, inside your network findings only, no rows sent

Checked N of about M rows: How many rows were read, and the estimated total.
Severity: kritisch, warn of informatie. A check shared by many columns collapses into one line naming the first few.
Issue counts: The tally by severity over the full list.
Column details: Each column with only its findings kept, plus the full profile, so your AI client can explain each issue.

What it does not help with

It does not fix anything. Nothing writes to your database.
It reports patterns, not causes. A failed leading-digit check is a reason to look, not proof of fraud.
Duplicate keys need a key: the table’s primary key on a plain read of one table, or the columns you name.
Like a profile, it reads every row it covers, so a slow query is still slow.

Grenzen

Up to 1,000,000 rows by default, at most 10,000,000; use a filter or a row limit on large tables.
The report is sized to fit the call’s budget. It sheds detail in a fixed order, never below the counts and the top ten issues, and says what it left out.
A quality report returns no rows, so your plan’s rows-per-query limit does not apply to it.

Instellingen

Your AI client sets the key columns, amount columns and rows to read on each call. The profile half uses your organization’s profile defaults. The other Deeper analysis defaults live on the Deeper analysis card on the Context tab.

When your AI client says no

When the server refuses, your AI client receives a code and one sentence, and usually relays it. Quoted sentences are the server’s own words; the parts in italics stand for the ids a real message names.

Code	What the server says	Wat te doen
`MCP_FEATURE_NOT_INCLUDED`	“Your plan does not include advanced analytics; anomalies, segments, comparisons and result diffs are available on Business Core.”	Ask your AI client to work from a profile, a follow-up query or rows instead, or move up to Business Core.
`MCP_AGENT_VERSION_TOO_OLD`	“Update the Network Agent to run data-quality checks at the source; this agent cannot answer mode=quality yet, so ask for a profile or rows until it is updated.”	Agents update themselves. Check the agent under Gegevens → Gegevensagenten.
`MCP_AGENT_NIET_VERBONDEN`	The Network Agent that owns the connector is not connected right now.	Open Gegevens → Gegevensagenten and start the agent on its computer.

Voorbeelden

“Can I trust the customer list for the mailing?”

Your AI client finds a share of emails in an invalid format and a batch of duplicate customer numbers, lists examples of each, and suggests a filter for the mailing.

“Do these supplier invoices look normal?”

The leading-digit check flags the amount column and the outlier check reports invoices far above the usual range. The client lists them for review.

Zit je nog steeds vast? Open Ondersteuning Ga naar Query Streams en laat ons weten in welke stap u zich bevindt en wat u ziet. Dan zorgen wij ervoor dat u verbonden wordt.

Facts source

Setup trouble?

Agent shows Offline, or Test connection and Browse are greyed out? Check that the Agent is online → Is the Agent online?