Resource

How to let an AI answer questions about your business without giving it guest data.

Every operator who considers AI eventually asks the same question. Does my guest data go to the model? Most vendors answer with a policy. The only honest answer is one you can test. Here is the architecture, and the test.

Pro Agentic StudioSeptember 20267 minute read

Written in September 2026 from our first build. In the platform today the model writes no queries: it asks for named views, and the guarantee below still holds.

An AI that can answer "how did direct bookings do last month compared to the same month last year" is genuinely useful to a busy operator. It is also the feature that makes a careful operator nervous, for good reason. To answer that question, the model has to read your data. The question is which data, and who decided.

What follows is how we built it. None of it is clever. All of it is checkable, which is the point.

01Policy is not architecture

The easy version of this feature reuses the existing read-only database role, the one the app's screens use, and adds a line to the AI's instructions telling it not to look at personal data.

That is a policy. It works until it does not. The read-only role can see the staging layer, the staging layer carries guest email addresses, and query results travel to an external model. The instruction not to look is the only thing standing between a guest's email and a third party. An instruction is not a wall.

An instruction is not a wall. The model cannot read what its database role cannot see, no matter what it is asked or how it is asked.

The right version makes the sensitive data unreachable by construction. Not forbidden. Unreachable. The model cannot read what its database role cannot see, no matter what it is asked or how it is asked.

02Two schemas, audited, nothing else

A warehouse for an operator has layers. Raw data lands as the source sent it. Staging types and keys it. Core and mart hold the cleaned dimensions, facts and display models that reports read from.

The AI role can read core and mart only. Before granting that, both schemas were audited column by column and confirmed to contain no email address, no phone number and no postal address. Guest identity lives in raw and staging, and the AI role is denied both, along with the operational tables and the system's own authentication tables.

Then the denial was verified by probe: log in as the AI role and try to read each forbidden schema. Every attempt fails. That result is recorded.

The architecture

The AI role can reach two schemas, and both were audited.

PipelineBIAIrawas the source sent itreaddenieddeniedstagingholds guest emailreadreaddeniedcoreaudited, no PIIreadreadreadmartaudited, no PIIreadreadreadopsruns and tokensreaddenieddeniedauthsystem tablesdenieddenieddeniedverified by probe, and a test fails the build if it changes
ReadableReadable by the AI roleDenied
Guest email addresses live in staging. The AI role is denied staging, raw, ops and the authentication tables outright, and every denial was verified by probe. It reads core and mart, which were audited column by column and contain no email address, phone number or postal address. That makes guest identity unreachable by construction rather than by instruction, which is the only version of this promise worth making.

Guest contact details never reach the AI. Not because it was told not to look. Because there is nothing there to see.

03A test that fails the build

An audit is a snapshot. The risk is what happens six months later, when someone adds a convenient email column to a mart table for a report, and the guarantee silently stops being true.

So there is a test. It scans every column the AI role can reach and fails the build if any of them looks like contact data. The first person to add an email column to the wrong place finds out before the change ships, not after a guest does.

This is the part to ask any vendor about. Not "do you protect guest data" but "what breaks if someone adds an email column to the tables your AI reads?" If the answer is a person noticing, the guarantee is a hope.

Schemas the AI can read
2
core and mart, both audited column by column
Schemas denied
4
raw, staging, ops and the authentication tables
Tests guarding it
1
the build fails if contact data ever reaches those two

04The query guard

The model writes database queries. Those queries need a fence around them, separate from the schema permissions above.

One detail, because everyone writes this bug first: blank out string literals before scanning for forbidden keywords. Otherwise a perfectly legitimate where status = 'create' is rejected because the scanner sees a write verb inside a piece of data.

The guard

Four gates, and the one everybody gets wrong.

Questionthe inputOne statementno chainingRead verbs onlySELECT or WITHRow capapplied by wrappingRole timeoutthe backstopevery query the model writesBlank out string literals before scanning for keywords, or a legitimatewhere status = ‘create’ is rejected as if it were a write. Everyone writes that bug first.
The schema permissions decide what the model can reach; the guard decides what it may run. The row cap is applied by wrapping the query, never by appending a limit to the end, because appending is trivially defeated and breaks on queries that already have one.

05Three smaller things that matter more than they sound

Read the catalogue at runtime. The list of tables and columns handed to the model is built from the database's own catalogue when the system starts, never by hand. A new view becomes answerable the moment it exists, and the model's picture of the database cannot drift from the real one.

One write, explicitly bounded. Every question and every query the model generates is logged, which needs a write, and the role is read-only. That one write is opened with an explicit, transaction-level escalation, and then tested to confirm it grants nothing else: the role still cannot create tables, and it cannot delete from its own log. Test the boundary. Do not assume it.

Hand the model the local date. A model given a date in universal time told an executive it was Sunday the 14th when Sunday was the 13th. Small, and it undermines everything else in the answer. The model gets the local business date, and the house style rules live in its instructions and are checked in its output.

06What we have not built yet

Honesty cuts both ways. The piece that is designed and not yet built is the feedback loop: a helpful or not-helpful control on each answer, captured against the question, the generated query and the shape of the result.

It matters because it is the only mechanism that turns use into improvement. It is how you learn which questions the data layer cannot yet answer well. It should be in the first version of any product like this, and in ours it is next.

The questions to ask any vendor

A vendor who has built this properly will enjoy those questions. That is a useful signal in itself.

Want to see the boundary tested live?

Thirty minutes. We will show you the AI answering a question about a sample portfolio, and then show you it failing to read anything it should not.