Your Tech Club
Book a call
INTEGRATIONSCHATGPT

Connect ChatGPT to your own systems

Answering a question about a customer, an order or a contract without opening five systems. That only works once ChatGPT can reach your own data. Here you'll find the routes available, what ChatGPT can read or do through them, and what happens to your data.

All integrations
4.8 from 127 reviews
ISO 9001 & 27001
ChatGPTAI assistants · OpenAI
  • Documents and files goes back and forth with Your other systems
  • Email and calendar goes back and forth with Your other systems
  • Customer and sales data goes back and forth with Your other systems
  • Financial and operational figures goes back and forth with Your other systems
Your other systemsSuch as Exact Online, AFAS Profit, HubSpot

Built for clients including Tata Steel, M&G Investments, Digistore24, Toppers in Concert, Bar-le-Duc and Loogman

What you exchange

This is the data organisations most often let flow between systems. Which of it matters to you depends on where work is getting stuck today.

Documents and files

Contracts, quotes, manuals and minutes in Google Drive, SharePoint or OneDrive. ChatGPT finds the right file and summarises what you need, within the permissions of the person asking.

Email and calendar

Gmail and Outlook: find an email thread, prepare for a meeting or have a draft reply ready. ChatGPT writes the draft, you decide whether it gets sent.

Customer and sales data

Contacts, deals and call notes from a CRM such as HubSpot or Salesforce. Useful for checking a deal's status or preparing for a customer call with the full history to hand.

Financial and operational figures

Outstanding invoices, revenue per customer, stock levels. Have ChatGPT pull these figures from the system rather than state them itself, and ask for the record or report behind them.

Popular integrations

For each combination: which data moves back and forth, what the standard integration does and what to watch out for.

Off the shelf or custom

Not always. If an existing integration for ChatGPT does exactly what you need, we'll tell you.

What already exists

For the common office suites and a growing number of business systems, there's an app you can simply switch on: Google Drive, Gmail, Google Calendar, SharePoint, Outlook, Teams and HubSpot are examples. Vendors such as Salesforce and Moneybird also offer their own MCP server, which you add as an app. For searching, summarising and asking questions, that's often enough.

On Business, Enterprise and Edu, ChatGPT uses Company Knowledge to search several connected sources at once. For Google Drive, SharePoint and Teams, an admin can also set up an indexed source; staff only find what they're allowed to see in the source system.

A standard app falls short when your system doesn't have one, when you need actions the app doesn't offer, or when the result has to land in a fixed process rather than a chat. We can connect any system, even when there's no standard integration.

When you need custom work
  • Your system has no app or MCP server, for example an in-house database or an industry-specific package.
  • ChatGPT also needs to do something in the system, and the existing app can only read.
  • The assistant has to follow business rules the standard integration skips, such as validations or an approval step.
  • You want the assistant in your own portal or workflow, not in ChatGPT itself.
  • You need more logging or narrower access than the app provides.
How it works technically

Apps in ChatGPT

The fastest route. There are ready-made apps (formerly connectors) for Google Drive, Gmail, SharePoint, Outlook, Teams and HubSpot, among others. You sign in to the system and ChatGPT works with whatever your account can see. In ChatGPT Business, apps are on by default; in Enterprise, a workspace starts with a selection and the admin can decide per role who uses which app.

Your own MCP app

If your system has no app, you build an MCP server (Model Context Protocol) that exposes that system's actions. An admin tests it in developer mode and publishes it to the workspace. Write actions over MCP are in beta for Business, Enterprise and Edu, on the web. For a server inside a closed network, there's Secure MCP Tunnel.

Through the OpenAI API

If you're building the assistant into your own application, you use the Responses API. You give the model a remote MCP server as a tool, or you define functions that your own code runs. By default, the API asks for approval before any data goes to an MCP server; you can switch that step off per tool.

Read, write and approve

Admins switch read and write actions on or off per app, where the app supports it, and choose whether new actions are added automatically. You also decide when ChatGPT asks first: always, only for changes, or only for higher-risk actions. Letting everything through without asking is only possible per app, not for the whole workspace.

Privacy and data residency

By default, OpenAI doesn't train on data from ChatGPT Business, Enterprise or the API, including what comes in through apps. You can sign a data processing addendum (DPA). Storage in Europe is available for new Enterprise customers and approved API customers. Data sent to a third-party app or MCP server falls under that party's terms.

AI in between

An integration moves data. In between, AI can read, sort and prepare a suggestion. Where it matters, a person decides.

Reading is not the same as writing

A summary of a customer file can't do much damage. A changed deal stage, a sent email or an edited document can. Start with read-only and switch on write actions one at a time, once the benefit is clear.

Drafts, not decisions

Let ChatGPT do the preparation: a draft reply, a suggested follow-up, a draft invoice. A person checks it and sends it. ChatGPT already blocks some high-risk actions itself.

Figures from the source

ChatGPT can make mistakes, including with numbers. Have amounts and quantities come straight from the system and ask for the record number or a link, so you can check.

Where it goes wrong

Permissions that are too broad

ChatGPT sees what the connected account sees. If a SharePoint folder or CRM role is too open, ChatGPT will find that data too. Tidy up permissions in the source system before you connect.

MCP servers you don't know

A custom or third-party MCP app isn't verified by OpenAI. A malicious server can send hidden instructions (prompt injection) or exfiltrate data. Only connect servers you trust, and review the tools before you publish.

Tools that change

Once published, ChatGPT uses a fixed snapshot of the tools. If the developer changes the server, calls can fail until an admin reviews the change. The admin doesn't get a notification about it.

Assuming everything stays in Europe

Storage in Europe applies to content in ChatGPT itself. Whatever goes to an app or MCP server follows that party's terms and location. So look at the processor behind every integration.

How we build an integration

  1. What goes where

    Which data moves, in which direction, and which system wins when two values differ. That is written down before anything gets built. For this integration, that means: Documents and files, Email and calendar, Customer and sales data, Financial and operational figures.

  2. Fields and rules

    For every field, how it translates between the systems: statuses, VAT codes, general ledger accounts, item numbers. And what happens with a duplicate or a missing value. For this integration, that means: Apps in ChatGPT, Your own MCP app, Through the OpenAI API, Read, write and approve, Privacy and data residency.

  3. Testing with real cases

    In a sandbox or test administration, with the exceptions you actually run into: credit notes, partial deliveries, changed customer details.

  4. Live, with logging and monitoring

    Every sync is logged, and when something fails, someone gets an alert. If a vendor changes its API, we update the integration.

The Your Tech Club team at the office in Amsterdam
Amsterdam
Who you'll talk to

No account manager. You talk to someone who builds integrations themselves, and who will tell you when a standard integration is enough.

Joey van VlietJoey van VlietFounder
Yassine SeghrouchniYassine SeghrouchniFounder
  • Office on the Duivendrechtsekade in Amsterdam
  • ISO 9001 and ISO 27001, certified by Kiwa
  • 500 projects delivered
  • 4.8 from 127 reviews
Office +31 (0)20 308 6200Your Tech Club on LinkedIn
Frequently asked questions

Does OpenAI use our business data to train its models?

Not by default. OpenAI doesn't train on data from ChatGPT Business, Enterprise or the API unless you opt in, for example by sharing feedback. That also applies to information ChatGPT fetches through apps. On personal plans such as Plus and Pro it can happen, depending on the setting to improve the model.

Does our data stay in Europe?

That depends on the plan. New ChatGPT Enterprise customers can choose storage in Europe (the EEA and Switzerland), and eligible customers can also have model processing take place in Europe. For ChatGPT Business, a storage region is being rolled out in phases, and even then OpenAI temporarily keeps a copy of prompts and responses in the US for abuse monitoring. Through the API, approved customers can have projects processed in Europe.

Can we sign a data processing agreement with OpenAI?

Yes. OpenAI signs a Data Processing Addendum for ChatGPT Business, ChatGPT Enterprise and the API; you request it through OpenAI's form. It sets out how OpenAI processes personal data on your behalf. If you connect systems with a lot of personal data, or sensitive personal data, you may also need a DPIA.

Can you use ChatGPT to assess job applicants or customers?

Stricter rules apply there. The AI Act classes AI used for things like selecting job applicants or credit scoring as high-risk, with requirements for human oversight and logging, among others, from 2 December 2027. If you use ChatGPT with customers, people must know they're dealing with AI; those transparency rules have applied since August 2026.

What is MCP?

Model Context Protocol is an open standard that lets an AI assistant call tools and data from other systems. An MCP server describes which actions are available, such as 'find customer' or 'create task', and carries them out with the permissions you grant when connecting. ChatGPT, Claude and other assistants use the same standard, so one server can serve several assistants.

Related services

All integrationsAccounting, CRM, ecommerce, HR and AI assistantsWhat does an integration cost?Get an estimate in two minutesCustom softwareWhen an integration alone isn't enough
AI assistants

Sources

The technical facts on this page come from the vendors' official documentation, checked on 19 September 2026. Vendors change their APIs and limits; when we build an integration, we always check the current state.