Skip to content
Project management Working integration

Connect Airtable to AI: ask the base a question instead of filtering it

Airtable can be connected to AI. The assistant retrieves records from the tables of the base you point at, adds a record on request and updates existing records. The connector looks inside one base at a time, so whatever sits in another base stays completely outside it.

Airtable usually starts as a handy list and grows into the administration a whole team runs on: stock, requests, content planning, suppliers. Once it reaches that size, time starts going into the retrieving itself. Somebody opens the base, finds the right table, sets a filter and copies the result into a message. For a question that could be asked in one sentence. The connector takes that step away. You ask in plain language and the assistant looks inside the tables of the base attached to the connector. It pulls the rows that belong to the question and leaves the rest alone, so an answer stays short and readable. If something has to be added, it creates a record using the fields your team uses in that table. If something has to change, it updates a record it looked up in the same conversation, and it first shows which record that is and what will change. A base is shared work: whatever it writes, colleagues see immediately.

What the assistant can do inside a base

Retrieving records from a table

The assistant reads the records of the table the question belongs to, with the fields your team uses there. That answers questions about stock, requests or planning without anyone opening the base.

Only the rows you asked for

It narrows down to what the question contains, for instance only the open rows or only this month, and can take an existing view as its starting point. That keeps a whole table out of the answer.

Adding a record

On request it puts a new record in the table you point at, filled in according to the fields already in use there. It shows what it wants to add before it lands.

Updating a record

Adjusting a status, correcting a number or adding a note happens on a record it looked up first. It names the record it is about, so you can be sure the right row changes.

One base per connector

The connector is attached to a single base. If you work with several bases, you set up a connector per base, so data from different teams never runs together.

Where a base really saves time

Retail

Situation
The stock of seasonal items lives in a base and the shop floor asks all day what is still there.
What the agent does
The assistant retrieves the rows for the item in question and names the quantity and the location.
Result
An answer in the chat instead of a colleague having to open the base between customers.

E-commerce

Situation
Returns are tracked in a table and at the end of the week it has to be clear what is still open.
What the agent does
The assistant collects the open rows, groups them by reason and puts the oldest on top.
Result
A short list the team can act on straight away, without a detour through a spreadsheet.

Marketing

Situation
The content planning lives in a base and a new request comes in that somebody has to schedule.
What the agent does
On request the assistant adds the item with the date, the channel and the owner the way the team records it.
Result
The planning stays complete without anyone updating the base between two meetings.

Industry and manufacturing

Situation
Machine maintenance is tracked in a table and the planner wants to know which services fall due this month.
What the agent does
The assistant retrieves the rows for the current month and notes which ones have already been signed off.
Result
Next week's planning is settled in minutes instead of after a walk through the table.

Customer service

Situation
A customer asks about the status of a request the team tracks in a base.
What the agent does
The assistant looks the request up and reports where it stands, and on request updates the status once something has been agreed.
Result
The customer hears it immediately and the administration does not lag behind the facts.

Setting the connector up per base

  1. 1

    Access to one base

    The connector is switched on with access you issue yourself that reaches one base. You withdraw it in the same place.

  2. 2

    Agreeing tables and roles

    You decide which tables take part, which roles may use the assistant and whether it may only read or also add and update records.

  3. 3

    What it does and does not see

    It sees the tables inside that base and, within them, the rows that belong to the question. Another base falls entirely outside, even if somebody asks for it.

  4. 4

    What is recorded

    Every consultation and every change lands in the IntraGPT log. In Airtable itself a record keeps showing when it was created or adjusted.

Connected securely

One base, not your whole account

A base often holds the operational core of a team, from customer details to purchase prices. That is why the connector reaches no further than the base you point at.

Where the data lives: Nederland of de EU

Access stays yours

The access is issued by you and withdrawn in one move. IntraGPT holds no employee passwords.

Bounded to one base

Other bases in the same account stay out of reach. To keep a base genuinely out, you do not have to configure anything further.

Writing can be traced

An added or changed record sits in the log with the moment and the request that preceded it. On top of that the assistant first shows what it is about to write.

Processing on a server of our own

The data from your base is processed on our own server in the Netherlands. It does not go to an American model provider and is not used to train models.

Airtable and AI: frequently asked questions

Can I connect Airtable to ChatGPT?

ChatGPT has no managed connection to your base, and moving operational data into a public chatbot is not an option for most organisations. IntraGPT connects the base you point at, writes only on request and records every action.

Does IntraGPT work with Airtable?

Yes. Airtable is a working integration already in use at customers. The assistant retrieves records from the tables of a base, adds records and updates them. Which base and which tables those are is agreed during setup.

Does the assistant see all our bases?

No. A connector reaches one base. If you have several and all of them should take part, we set up a connector per base. That is also exactly how you keep the data of different teams apart.

Can the assistant create or change records?

Only if you switch that on, and only on an explicit request. It first shows which record it wants to add or change and what will go into it. If you would rather not, it only reads and nothing in the base ever changes.

Could the wrong row be overwritten by accident?

That is exactly why it looks a record up and names it before updating it. It does not edit a row it has not found in the same conversation, and when two similar rows could be meant it asks which one you want instead of guessing.

Our tables are set up our own way, does that work?

Yes. The assistant works with the fields as your team named them, not with a fixed layout of ours. It does help if those fields are filled in consistently; during setup we look together at the tables that will be used most.

What does an Airtable connector cost?

Connecting sits inside the project we run together. What drives the work is the number of bases and tables, whether writing is allowed and which roles may use the assistant. We work that out in the first conversation.

How long does setup take?

A connector on one base is up quickly. The time goes into walking through the tables that take part and agreeing whether writing is allowed. If more bases follow later, that is a small extension.

For which sectors

Other connectors

See all connectors

Curious what this connector would deliver for you?

Book a free thirty minute AI call. We look at your systems, the permissions around them and the first use case that saves time or money.

Book a free AI consultation