Intro
ChatGPT is no longer just an AI you chat with; it’s starting to behave like a remote control for your digital life. By wiring services like Spotify, Uber, DoorDash, Booking.com and others directly into the assistant, OpenAI is testing what happens when you stop opening apps and instead just describe what you want done. That shift sounds subtle, but it points to a much bigger play: turning ChatGPT into a cross‑app operating layer. In this piece, we’ll look at what TechCrunch’s report really signals — for users, for platforms, for regulators and especially for those of us in Europe who are, once again, watching from the sidelines.
2. The news in brief
According to TechCrunch, OpenAI has rolled out a catalogue of app integrations inside ChatGPT that let users connect third‑party accounts and trigger actions by natural‑language prompts. Once logged into ChatGPT, you can either type an app’s name at the start of a prompt or visit a new Apps and Connectors section in Settings to link services.
The current lineup spans home services (Angi), travel (Booking.com, Expedia, Zillow), design and productivity (Canva, Figma, Wix, Coursera, Quizlet), shopping and food (Target, DoorDash, Uber Eats), mobility (Uber) and media (Spotify). After connection, ChatGPT can, for example, assemble meal plans and push ingredients into DoorDash, create Spotify playlists, or draft a Wix site.
TechCrunch notes that more partners, including OpenTable, PayPal and Walmart, are planned for 2026. For now, the rollout is limited to the U.S. and Canada; users in Europe and the U.K. do not have access.
3. Why this matters
What looks like a quality‑of‑life upgrade is actually a platform land grab.
For users, the upside is obvious: instead of hopping between half a dozen apps and wrestling with filters and forms, you describe an outcome — “three‑day city break in April, direct flights only, hotel with breakfast and good Wi‑Fi” — and let the assistant orchestrate the rest. The friction reduction is real, and once people get used to this, going back to manual app‑hopping will feel archaic.
But there is a trade‑off: connecting an account means exposing behavioural data — playlists, purchase history, movement patterns, home search criteria — not only to the original provider but to OpenAI as an additional layer. That creates richer personalization, but also a much more detailed profile of your life, and a single point of failure if something goes wrong.
For OpenAI, these integrations are strategically priceless. They turn ChatGPT from a destination product into the interface through which you access other products. That is the difference between being an app on the phone and becoming the thing that makes apps feel optional. Once users habitually ask ChatGPT to get things done, OpenAI gains both stickiness and leverage over partners.
Partners like Spotify, Booking.com or Uber gain new distribution and a powerful conversational front‑end they didn’t have to build. Yet they also risk being commoditized into interchangeable back‑end utilities. If the user’s memory of the transaction is “I asked ChatGPT to handle it,” the brand in the background becomes less visible — and more replaceable.
4. The bigger picture
These integrations sit squarely inside a broader shift from static chatbots to AI agents that can take actions in the world. Google pitches Gemini as something that can book trips and manage documents; Microsoft’s Copilot is being wired deep into Windows and 365 to automate workflows; Meta wants its AI layer inside WhatsApp and Instagram. OpenAI’s move is the most explicit version of the same trend: the assistant as orchestrator of services.
There’s also a strong echo of the super‑app model that made WeChat dominant in China. Instead of mini‑programs, we now have connectors. Instead of tapping icons, we talk. In both cases, one interface mediates most daily digital tasks. The West has long failed to build its own super‑app because of platform silos and regulation; AI provides a plausible new route, because an assistant can sit above iOS, Android, web and desktop simultaneously.
Historically, whoever controlled the primary user interface — the browser, the mobile OS, the app store — captured outsized power and rent. App integrations are how OpenAI tests whether it can become that layer for the AI age. If ChatGPT is where intent is expressed and decisions are made, argument over which shopping or travel site ranks first on Google suddenly matters a lot less.
The comparison with earlier "skills" ecosystems, like Amazon Alexa’s, is instructive. Voice platforms struggled because they were bad at understanding complex, open‑ended requests and even worse at memory and context. Modern large language models drastically change that equation. The experience is still far from perfect, but it’s good enough that the convenience trade‑off starts to tilt.
5. The European / regional angle
The most immediate European angle is simple: we don’t get this — yet. TechCrunch explicitly states that app integrations are restricted to the U.S. and Canada, with Europe and the U.K. excluded. That’s unlikely to be a random rollout decision; the data‑sharing and profiling implied by these connectors sits right in the crosshairs of GDPR and the upcoming EU AI Act.
Once you let an AI layer read your travel, shopping, transport and media data, questions multiply: Who is the data controller — the app, OpenAI, or both? On what legal basis is data combined for new purposes? How is consent captured and withdrawn for each connector? How do you prevent covert profiling across domains? All of this is manageable, but it is not simple.
Add to that the Digital Markets Act (DMA). If ChatGPT, backed by Microsoft, ever ships as a deeply integrated assistant in Windows or as a default in major apps, it could trigger “gatekeeper” rules around self‑preferencing and access for rivals. The EU will be keen to avoid simply swapping one set of mobile gatekeepers (Apple/Google) for an AI gatekeeper that sits above them.
For European startups and incumbents, there is both risk and opportunity. Risk, because if a U.S. AI layer becomes the default interface for travel, commerce or media, local players may be reduced to invisible suppliers. Opportunity, because there is still space to build Europe‑first stacks — from foundation models (Mistral, Aleph Alpha, Stability) to privacy‑conscious assistants — that bake compliance and data residency in from day one.
6. Looking ahead
Three things are worth watching over the next 12–24 months.
First, how far OpenAI pushes into high‑risk domains like finance and health. Today’s partners are relatively low‑stakes from a regulatory standpoint. The real test of the agent model will be when users expect ChatGPT to move money, sign contracts or touch medical data. That’s where both compliance and liability become much more serious — especially under EU rules.
Second, the battle over default interfaces. Big tech vendors will not happily hand control of user intent to a third‑party assistant. Expect Google, Apple, Microsoft and Meta to double down on their own agent layers and to push them as defaults in their ecosystems. European regulators will simultaneously try to keep those defaults contestable under the DMA. The question becomes: will we end up with one dominant AI interface, or a patchwork of competing agents with partial integrations?
Third, the European rollout question. If and when OpenAI wants these connectors live in the EU, it will need a crisp story on consent, data minimisation, joint controllership and user redress. That almost certainly means a slower, more granular permission model than the U.S. version — and possibly region‑specific data handling. A realistic scenario is that full‑fat agent features arrive in Europe 12–18 months after North America, unless EU‑based competitors force a faster move.
For companies, the homework is clear: think about agent‑readiness. Are your APIs and data models structured for an AI to use safely? Can you expose capabilities without oversharing data? Do you have logging and audit trails for AI‑driven actions? The winners of the agent era will be those who quietly solve these unglamorous plumbing problems early.
7. The bottom line
ChatGPT’s new app integrations are less about convenience buttons for Spotify or Uber and more about testing whether an AI assistant can become the universal front‑end for digital life. If OpenAI succeeds, power will shift from app stores and search results to whichever agent users trust with their intent and their data. Europe has a narrow window to shape how — and under whose rules — that happens. The real question for users and policymakers alike is simple: who do you want orchestrating your day?
