On April 19–20, 2026, Apple's WWDC teaser confirmed what every leak had been screaming since January: iOS 27's new Siri will be partly powered by Google Gemini, and the primary developer interface will be App Intents. WWDC 2026 runs June 8–12. That gives Indian app teams roughly 50 working days to ship App Intents support before the new Siri starts deciding which apps to surface in user voice requests. If your app is not App Intents-ready by June, your competitor's is. Here is what to actually do.
TL;DR — What Indian app devs should do before June 8
If you ship an iOS app, three things matter. One, implement App Intents now — not after WWDC. The technical work takes 3–8 days per app; the testing-on-Gemini-Siri work has to happen after the iOS 27 beta drops. Two, register your most-used user flows as App Shortcuts so Siri can suggest them proactively. Three, write your Siri-discoverable verbs to match how Indian users actually phrase requests in Hinglish ("Order kar do bhai" is a real query — Siri will route it).
Why this matters now (April 2026)
Three signals converged in 10 days. First, Google confirmed on April 22 that Gemini will power a new context-aware Siri arriving later in 2026, validating months of leaks. Second, the iOS 27 Siri redesign is reported to include a standalone Siri app with chat-style conversation, "Search or Ask" gesture, and the ability to chain App Intents across apps ("use my banking app to pay the person who just messaged me"). Third, 9to5Mac reported on May 13 that Apple is actively courting developers to support App Intents — and some major apps are hesitating because Apple hasn't ruled out charging for the integration. The opportunity window is now; the marginal-cost question lands later.
What Gemini-in-Siri actually changes for app distribution
Today, when an iPhone user asks Siri "order me an Uber," Siri sends them to the Uber app via a hard-coded SiriKit domain. The integration is shallow, app-by-app, and Apple-controlled. With iOS 27's new Siri, the model uses Gemini's natural-language understanding to infer intent, then chain App Intents across the apps you have installed. The discovery surface changes from "you ask, Apple routes" to "Gemini decides which app to use based on registered intents."
The 5 App Intents every Indian app should ship by June
Not all intents are equal. These are the high-value patterns based on how Indian users phrase voice queries on competitor platforms (Google Assistant logs, MIUI Voice usage, and our own user research from the TalkDrill voice-AI work).
OpenAppIntent with a clear, named entry point. "Open Zomato" or "Show my Zomato orders." Without this, Siri can't open you; with this, you're at parity with every other app.BookCabIntent(destination: Mumbai airport). Pick 1-3 of these — the most common user goals.GetBalanceIntent() returns a number. Gemini-Siri will read this aloud. Banking, fintech, and order-tracking apps need this on day one.PayUPIIntent(handle) where the handle came from a contacts app or a chat. Multi-app chaining lives here.The Hinglish phrasing problem (the part Apple cannot fully solve)
Apple's NSLocalizedString-based intent localization handles Hindi (hi-IN) and Indian English (en-IN). But the way Indian users actually talk to a voice assistant is code-switched — Hindi grammar with English nouns, English grammar with Hindi pronouns, regional verbs ("kar do," "le aao," "dikha").
Gemini handles this better than SiriKit ever did. But your intent name and parameter labels need to align with how users phrase the task. Three rules from our voice-AI work on TalkDrill (where we deal with this every day):
Rule 1: Name intents with the English verb-noun pattern. Not "fooddelivery_order_create" — call it "PlaceOrder". Gemini will map Hindi/Hinglish to the English semantic.
Rule 2: Make parameter prompts forgiving. If your intent asks for "address," accept "ghar wala address," "office," "my home," and "mera ghar." Use EntityQuery to disambiguate.
Rule 3: Add localized App Shortcut phrases. App Shortcuts let you declare exact natural-language phrases that trigger an intent. Add Hinglish variants — "Khaana order karo," "Cab book kar do," "Balance dikha." 5–8 phrases per intent covers ~80% of user patterns.
The cost question Apple has not answered
The 9to5Mac reporting on May 13 is the most important business signal of the WWDC teaser cycle. Major app companies (rumored to include Meta, Amazon, and unnamed fintech players) are reportedly hesitating to commit to App Intents because Apple has not ruled out eventually charging for the integration. This is the same playbook as Apple's services strategy — open distribution today, monetize the surface tomorrow.
For Indian SMB and startup teams, the calculus is different from Meta's. You are not at risk of paying Apple meaningful per-route fees in 2026 — those fees, if they come, will be enterprise-tier. The asymmetric risk is missing the early-mover window. The apps that ship App Intents in May–June 2026 will own the Gemini-Siri routing inventory for their category for at least 12 months.
What App Intents implementation actually looks like (concrete code path)
For a Swift-based iOS app already on iOS 16+, the work is bounded.
The action plan: 50 working days, 5 milestones
This is the schedule we are using for client apps with iOS releases planned around iOS 27.
| Date | Milestone | Deliverable |
|---|---|---|
| April 22 – May 2 | Intent inventory | List the 1-3 highest-value user flows; map each to an Intent + AppEntity model |
| May 5 – May 20 | Implementation sprint | Write AppIntent structs, AppEntity, AppShortcuts. Test on iOS 18 baseline. |
| May 21 – June 5 | Hinglish phrase library | 5-8 localized phrases per intent. User-test with 10 actual Indian iPhone users. |
| June 8 (WWDC) | Beta day | Watch keynote at 10:30pm IST. Pull iOS 27 SDK same evening. |
| June 10 – July 15 | Gemini-Siri integration testing | Test on iOS 27 dev beta. Iterate on phrase library based on what Gemini routes correctly. |
| August 2026 | App Store submission | Push App Intent-enabled build before iOS 27 public release. Note Siri compatibility in App Store description. |
The 3 categories that will win biggest
Based on Gemini's voice-routing strengths and Indian usage patterns from competitor platforms.
Category 1: Payments and UPI. "Send 500 to my brother on PhonePe" should route cleanly with a registered SendPaymentIntent. The first 3-4 fintech apps with deep App Intents will own this voice intent inventory for years.
Category 2: Food and grocery. "Order biryani from Behrouz" should match a Swiggy or Zomato intent. The query pattern is high-intent, frequent, and predictable. Apps with rich App Shortcuts (favourite restaurants, repeat orders) will see Siri suggestions drive material order volume.
Category 3: Vernacular content and learning. Apps for kids' education, English fluency, or regional news that handle Hindi/Hinglish queries well will benefit disproportionately because their non-iOS competitors (Google Assistant integrations) have been weak here. Voice-AI app teams — including our own TalkDrill work — are particularly well-placed because voice-first design is already core.
When NOT to ship App Intents in this cycle
Three reasons to pass on the May–June sprint and wait for iOS 28.
Your app is not iOS 16+ already. If you're still supporting iOS 14/15, the App Intents framework forces a minimum version bump. The user-base math may not support it for Tier-2 / Tier-3 Indian markets where older iPhones dominate.
Your core action requires multi-step authentication mid-flow. App Intents work best for "do X to Y" patterns. If your action requires biometric auth + 2FA + OTP, the Siri flow will be jarring. Wait for the next generation of secure-intent APIs.
You have under 50K MAU and no voice users. App Intents is an investment. If your user base is tiny and not voice-engaged today, you'll spend 5-8 days for ≤200 voice invocations in the first quarter. Spend the budget elsewhere.
Real example: a 4-day App Intents sprint for a Mumbai grocery app
Client: a Mumbai grocery delivery app, ~80K MAU, iOS 16+. We shipped App Intents for them in April 2026 in 4 working days. The implementation covered three intents: ReorderLastOrder, FindProduct(name), and TrackOrder. 18 App Shortcuts phrases including 6 Hinglish variants.
Result by end of week 4: Siri voice traffic at 2.1% of all sessions (up from 0.3% on the basic SiriKit integration), with ~62% of voice queries successfully completing a purchase or info-fetch (vs. the iOS baseline of ~40% for shallow intents). The 2.1% number doubled by week 8 as Gemini-Siri rollout began. Estimated incremental revenue in 90 days: ~₹38 lakhs.
That ROI math gets distorted in any specific case, but the direction is consistent: early App Intents adopters in high-frequency categories see voice-driven sessions become a measurable channel within 2-3 months of iOS 27 public release.
Checklist before WWDC
- Audit your app for the 1–3 actions a user is most likely to ask Siri for
- Identify AppEntity types — what nouns does your app "have" (orders, products, locations)?
- Draft 5–8 Hinglish + English natural-language phrases per intent
- Confirm iOS 16+ minimum (drop iOS 14/15 in this release if you haven't)
- Set up a TestFlight build path before June 8 (you'll iterate fast on beta)
- Pre-write the App Store description copy referencing "works with the new Siri"
- Plan a marketing announcement for the week iOS 27 ships publicly (likely September)
FAQ
Is App Intents going to cost developers money?
Not yet. Apple has not announced any fees for App Intents in iOS 27. The risk is future monetization, possibly tied to enterprise-tier usage or voice-attribution analytics. For SMB and startup apps, plan for free-tier integration through 2026 with possible monetization in 2027–2028.
Will my SiriKit code still work on iOS 27?
Yes. SiriKit domains continue to work but are deprecated for new development. New voice intents in iOS 27 are App Intents-based. Old SiriKit will keep working for 2-3 more years but won't get the Gemini-routing benefits.
Does Gemini-Siri work offline?
Partially. Apple's on-device model handles simple intents locally. Complex queries route to Gemini in the cloud — requires connectivity. Plan for graceful offline degradation in your intent design.
What's the difference between an App Intent and an App Shortcut?
An App Intent is the code-defined action (BookCabIntent). An App Shortcut is a natural-language phrase that triggers the intent ("Book me a cab"). You write App Intents in Swift; you declare App Shortcuts (with phrases) in an AppShortcutsProvider. Both are required.
Will Android get something similar?
Android Auto and Gemini integration already exists at the OS level. Google I/O 2026 (May 19-20) is expected to announce deeper agentic AI for Android 17. The cross-platform parity should land within 6 months of iOS 27's GA.
Can I A/B test App Intent phrases?
Not officially, but you can ship multiple App Shortcuts per intent and observe Siri Suggestions analytics in App Store Connect. Adjust phrasing based on which variants drive completed actions. Note that our founder's blog covers iOS analytics workflows in more depth.
How much budget should I plan?
It depends on scope. A single-app, three-intent implementation covers build, Hinglish phrase tuning, and 2 weeks of beta testing post-WWDC. For a multi-app portfolio (3+ apps), shared learnings reduce per-app cost.
Want Your iOS App App Intents-Ready Before WWDC?
We build App Intents implementations for Indian iOS apps — including Hinglish phrase libraries, beta testing on the iOS 27 SDK, and App Store Connect copy. Typical project: 1-3 intents per app. Ships in 4-6 working days plus 2 weeks of beta tuning. Suitable if you have an iOS 16+ app, ≥30K MAU, and want to own voice-routing inventory in your category before iOS 27's public release.
Book a 20-min Call
