Whether you're looking for an angel investor, a growth advisor, or just want to connect — I'm always open to great ideas.
Get in TouchOpenAI adds native DoorDash, Spotify, and Uber integrations to ChatGPT, letting users order food, play music, and book rides without leaving the chat — a decisive move toward building the AI-native super-app.
AI, startups & growth insights. No spam.
TL;DR: OpenAI has launched native integrations for DoorDash, Spotify, Uber, Canva, Figma, and Expedia directly inside ChatGPT. Users can now order food, queue music, book rides, and design visuals without leaving the chat interface. The feature connects existing accounts via OAuth and is available to ChatGPT Plus subscribers. The move signals OpenAI's ambition to turn ChatGPT from a question-answering tool into an AI-native operating system for everyday life — a model ripped straight from WeChat's super-app playbook.
On March 14, 2026, TechCrunch reported that OpenAI had quietly launched a suite of third-party app integrations inside ChatGPT. The integrations cover six services at launch: DoorDash, Spotify, Uber, Canva, Figma, and Expedia.
The feature is available to ChatGPT Plus subscribers and works by connecting your existing accounts on those platforms to ChatGPT via OAuth. Once connected, ChatGPT can take actions on your behalf — placing a food order, controlling music playback, requesting a ride — without you ever leaving the conversation.
This is not a chatbot pointing you toward another app. The actions happen inside ChatGPT. The conversation interface becomes the interface.
The timing is deliberate. OpenAI CEO Sam Altman has been explicit about targeting 1 billion weekly active users as the next milestone. ChatGPT currently sits at roughly 900 million weekly active users as of early 2026, according to a16z's sixth edition Top 100 Gen AI Apps report. Launching transactional integrations at this scale transforms ChatGPT from a productivity tool into a daily-life platform — and changes the revenue calculus entirely.
The architecture is straightforward: ChatGPT uses OAuth 2.0 to connect to partner service APIs. When you link a service, you authorize ChatGPT to act as a client on your behalf, scoped to specific permissions (read your order history, create a new order, control playback, request a ride).
Inside a conversation, ChatGPT identifies intent from natural language — "order my usual from Chipotle" or "play something like Tame Impala" — and maps that to a structured API call to the partner service. The response comes back into the chat thread as a confirmation, a receipt, or a status update.
Three architectural details matter here:
Tool calling. OpenAI's function-calling framework, first introduced in 2023 and refined through the GPT-4 and o-series model updates, is what makes this possible. When ChatGPT receives a request it can route to a connected integration, it invokes the appropriate tool function with structured parameters. The model decides which tool to call; the tool executes the real-world action.
Scoped permissions. Each integration requests only the permissions it needs. DoorDash needs your delivery address and payment method on file. Spotify needs playback control and library access. Uber needs your location and saved payment. Users can revoke permissions at any time from ChatGPT's settings.
Confirmation gates. For transactional actions (ordering food, booking a ride), ChatGPT shows a confirmation step before executing. You see what it's about to do, the price, and the details. You approve it. This matters both for user trust and for liability — OpenAI is not the merchant of record on these transactions.
The integrations are currently limited to the ChatGPT web app and iOS. Android support is listed as coming soon.
The DoorDash integration is probably the most immediately useful of the six. Link your DoorDash account and ChatGPT gains access to your order history, saved addresses, and payment methods already on file in the app.
You can ask ChatGPT to reorder a past meal ("get me that Thai order from last Thursday"), browse a restaurant menu while describing dietary constraints, or get a recommendation based on what you're in the mood for. ChatGPT will suggest restaurants, walk you through the menu, handle item customization, and place the order — all in conversation.
The practical workflow looks like this: you open ChatGPT, say "I want Thai food delivered to my apartment, nothing too spicy, under $20," and ChatGPT browses nearby DoorDash restaurants matching those criteria, presents two or three options with estimated delivery times, asks you to confirm, and places the order. You get DoorDash's standard order tracking from that point forward.
The integration also supports scheduled orders. "Order me a coffee from Blue Bottle for 8:30 AM tomorrow" is a valid instruction that ChatGPT can handle end-to-end.
What it cannot do yet: apply promo codes not already in your account, split orders across multiple restaurants, or handle DashPass subscription management. These limitations are consistent with a v1 integration scoped to core ordering flows rather than account management.
The Spotify integration spans two use cases: playback control and music discovery.
On the control side, ChatGPT can play, pause, skip, adjust volume, shuffle, and queue tracks — the same actions available through Spotify Connect. If you're mid-conversation and want to change what's playing, you can do it in natural language without switching apps. "Turn this down a bit" or "skip to something more energetic" work as you'd expect.
The discovery side is where the integration becomes genuinely interesting. ChatGPT can search Spotify's catalog and build playlists based on natural language descriptions that Spotify's own search would struggle to interpret. "Make me a playlist for a 45-minute run that builds intensity in the last 15 minutes" is the kind of request that fits conversational AI far better than a search box.
ChatGPT can also cross-reference your listening history to personalize recommendations. If you ask for "something like what I've been listening to lately but for cooking dinner," it has enough context from your Spotify library to do that meaningfully, rather than defaulting to a generic recommendation.
The integration requires a Spotify Premium account. Free tier users can see recommendations and search results, but playback control is restricted to Premium — consistent with Spotify's existing limitations on its own API.
One notable capability: ChatGPT can add songs or playlists it recommends directly to your Spotify library. You hear about a track in conversation, say "add that," and it's in your library. The friction between discovery and saving is effectively eliminated.
The Uber integration handles ride requests, fare estimates, and booking from within ChatGPT. Link your Uber account and ChatGPT gains access to your saved addresses, preferred ride types, and payment methods.
The basic flow: "Book me an Uber to SFO, I need to be there by 2 PM." ChatGPT checks current Uber pricing and estimated pickup times, recommends a ride type (UberX, Uber Comfort, Uber Black) based on your history and stated preference, shows the fare estimate, and books the ride on confirmation. You get a standard Uber push notification and can track the driver in the Uber app.
ChatGPT handles the conversational complexity that Uber's own interface doesn't. "I have a meeting at 315 Park Ave at 3 PM, when should I leave from here?" ChatGPT can factor in current traffic conditions via Uber's API, current surge pricing, and your stated address to suggest a departure time and pre-book accordingly.
The integration also supports Uber Eats, the food delivery side of Uber. This creates an interesting overlap with DoorDash — you have two competing food delivery integrations in the same interface, and ChatGPT can theoretically compare pricing between them before choosing where to route your order. OpenAI has not confirmed whether multi-service comparison is an intended use case, but it's technically possible given the current architecture.
Scheduled rides work similarly to DoorDash's scheduled orders. "Book me an airport pickup for 6 AM on Monday" is a valid instruction ChatGPT can execute.
The three non-consumer-services integrations round out the launch set and reveal a second prong to OpenAI's strategy: creative and productivity tools.
Canva lets ChatGPT create and edit designs directly. Describe a social media post, a presentation slide, or a simple graphic and ChatGPT will generate it in Canva, opening the result for you to fine-tune. This is a natural extension of ChatGPT's image generation capabilities, now with Canva's templating, brand kits, and export options behind it.
Figma is the developer-facing version of the same idea. ChatGPT can read Figma files (with permission), discuss design decisions, generate design tokens, and create new frames based on specifications described in conversation. For design teams that already use ChatGPT for writing specs and documentation, this closes the loop into the design tool itself.
Expedia mirrors the DoorDash and Uber pattern but for travel. ChatGPT can search flights, compare hotel options, build itineraries, and book travel — all against your Expedia account and preferences. "Plan me a long weekend in Lisbon in April, budget flights, boutique hotel, nothing too touristy" is the kind of open-ended travel brief that Expedia's own search interface handles poorly and that conversational AI handles well.
These integrations are not a standalone product decision. They are a strategic declaration.
Sam Altman has discussed reaching 1 billion weekly active users as an explicit company goal. ChatGPT is at 900 million. The remaining 100 million is not a content or capability gap — it's a habit gap. The people who haven't made ChatGPT a daily-use tool are largely people for whom it has not yet been integrated into their actual daily routines.
Ordering food, booking a ride, and queueing music are not activities that require AI intelligence. They are activities that require only a familiar interface and low friction. By becoming the place where you do these things, OpenAI gets what every super-app has always been after: daily, habitual engagement that isn't tied to a specific task.
The strategic logic is identical to what made WeChat dominant in China. WeChat started as a messaging app and gradually absorbed payments, food delivery, ride hailing, appointments, games, and mini-programs until it became the primary interface through which hundreds of millions of people interact with digital services. It is not the best app for any single task. It is the app people open first, because everything is already there.
OpenAI is attempting the same pattern, with two significant structural advantages WeChat didn't have: a general-purpose conversational AI at the center, and a global distribution base that WeChat never achieved outside China.
The WeChat analogy is useful but imprecise. It is worth examining where it holds and where it doesn't.
WeChat's super-app status in China was enabled by a few specific conditions: a dominant messaging platform that made itself the communication layer, a captive developer ecosystem through WeChat Mini Programs, regulatory conditions that limited Western competition, and a payments infrastructure (WeChat Pay) that made every transaction sticky. Remove any one of those and WeChat's dominance looks more fragile.
ChatGPT has none of these exact conditions. It does not have a dominant messaging layer in the Western market — iMessage, WhatsApp, and SMS all sit between ChatGPT and daily communication. It does not yet have a payments layer, though the integration architecture clearly points in that direction. And it faces vigorous competition from Google (Gemini), Anthropic (Claude), and Meta (Llama-based products), none of which WeChat had to contend with.
What ChatGPT does have is something WeChat lacked: a genuinely useful AI at its core. WeChat's mini-programs succeeded despite WeChat's AI being non-existent; they succeeded because of distribution. ChatGPT's integrations can succeed because the AI adds real value to the task. Natural language ordering, context-aware recommendations, and multi-step orchestration across services are things the AI does better than a dedicated app, not just differently.
The bet OpenAI is making is that an AI-native interface, where the model understands context, executes across services, and learns from each interaction, is structurally superior to the app-grid model of the smartphone era. If that bet is right, the comparison to WeChat undersells what ChatGPT is becoming.
OpenAI has not publicly disclosed the financial terms of these integration partnerships. The most likely revenue architecture has three layers:
Referral or transaction fees. Standard practice for platform-to-service integrations. A percentage of each DoorDash order, Uber fare, or Expedia booking that originates through ChatGPT flows back to OpenAI. This is how Google earns on hotel and flight searches, how Apple earns on App Store in-app purchases, and how WeChat Pay earns on WeChat-initiated transactions.
Premium tier lock-in. The integrations are currently Plus-only. This strengthens the value proposition for the $20/month subscription, reducing churn among subscribers who would otherwise drift back to free tier. As more integrations launch and the behavior becomes habitual, switching costs rise.
Data and personalization loop. The more transactions users run through ChatGPT, the richer the behavioral model becomes. A user whose ChatGPT account connects to DoorDash, Spotify, Uber, Expedia, Canva, and their calendar creates a cross-service profile that no individual app can replicate. This has downstream value for personalization quality, which in turn drives engagement.
For the partner companies, the calculus is different. DoorDash, Uber, and Spotify all benefit from any new acquisition channel that drives orders or streams. But they also face channel risk: the more dependent they become on ChatGPT as a distribution layer, the more leverage OpenAI accumulates over their customer relationships. This is the classic platform dependency tension that app developers have always faced with Apple and Google, now showing up in AI.
For Canva and Figma specifically, the dynamic is more collaborative. ChatGPT generates content; Canva and Figma provide professional tools to refine and export it. The integration makes both tools more useful together than either is alone.
The implications for app developers, product teams, and platform builders extend well beyond the six integrations in this initial launch.
The standalone app is under new pressure. Apps built around a single, well-defined task — order food, book a ride, stream music — are the most vulnerable to integration into conversational AI. If ChatGPT can do what DoorDash's app does, with less friction, DoorDash's direct relationship with the user weakens. The brand survives, the service survives, but the app as the primary user interface becomes optional.
Aggregators are at risk. Yelp, Booking.com, and similar aggregators built their businesses on being the discovery and comparison layer between users and services. ChatGPT with Expedia integration is a better version of Booking.com for most casual travel queries. As AI integrations expand, every aggregator business model deserves reexamination.
API-first companies benefit. The companies best positioned to get integrated into ChatGPT (and Google's Gemini, and Anthropic's Claude) are those with clean, well-documented public APIs. The era of AI super-apps will favor services that were built to be composable. Services built around proprietary apps with no external API access will be left behind.
The onboarding funnel flips. Today, a user discovers DoorDash, downloads the app, creates an account, adds payment details, and places a first order. With ChatGPT integration, the discovery and first order can happen entirely within ChatGPT before a user ever opens the DoorDash app. Acquisition now happens at the AI platform layer, not the app layer. Marketing budgets, app store optimization strategies, and user acquisition funnels built around direct app downloads need to account for this.
OpenAI becomes infrastructure. The most consequential reading of this launch is that OpenAI is evolving from an AI model provider into a platform layer — infrastructure that other businesses run on. This is a different business than building the best AI model. It is the business of owning the user relationship at the level where decisions happen. That is a much larger and more defensible position.
The six integrations that launched on March 14, 2026, are not the destination. They are the opening move in a platform strategy that, if it works, will reshape how applications are built, distributed, and monetized for the next decade.
Sources: TechCrunch — "How to use the new ChatGPT app integrations, including DoorDash, Spotify, Uber, and others" (March 14, 2026); a16z Top 100 Gen AI Consumer Apps, Sixth Edition (March 2026); OpenAI product blog
OpenAI is preparing to integrate its Sora AI video generator directly into ChatGPT after the standalone app saw installs plummet 45% in January — a major pivot that could bring video generation to hundreds of millions of users.
a16z top 100 gen AI apps sixth edition: ChatGPT hits 900M WAU, Claude grew 200%+ paid subscribers, Gemini leads growth rate.
Google embeds Gemini directly into Google Maps with a conversational 'Ask Maps' feature, giving 2B+ monthly users natural language search across 300M+ places — the biggest consumer AI deployment yet.