This roadmap is outdated and no longer reflects current plans.
It is preserved here for posterity as a snapshot of the V2 improvement thinking circa May 2026. Workstreams, sequencing, and dates have since moved on. For the current state of the product, see the Application Catalog. Treat everything below as historical.
V2, what's next*
The next chapter of V2. A set of improvements that, together, make the product feel like it's paying attention. Timed to land with the neural-network rebrand, so the product can demonstrate the intelligence the brand is claiming.
These dates are directional.
Nothing in this deck is a committed milestone. The dates are a high-level read on the sequencing we are working toward, anchored to the rebrand window. They will move as resourcing, scope, and parallel programs come into focus.
Treat the deck as a shape of the program, not a schedule. The June and early-July releases are the firmest. Anything past V2.5 is plausible sequence, not promise.
- V2.2, V2.3, V2.3.1 are already underway. They firm up first.
- V2.4 and V2.5 carry the V2 narrative. Scope is still being agreed.
- Jazz, Bancus, payment refactor, and YoMobile share the same engineering bench.
- Flights and insurance are deprioritized and blocked by external partners. They show as TBD on purpose.
The rebrand frames Global YO as a neural-network intelligence layer for travel. The product has to back that up where users actually see it.
- The home screen notices the user. The Active Main Hub surfaces what matters right now, and reorders as that changes.
- The system says what it noticed. Proactive Yolanda is the voice pattern. One nudge, one action, never an explanation.
- The app feels physically responsive. Navigation hints, haptics, polish. The product gives weight to every gesture.
Bars span work in flight to release. Grid lines are weeks. Grouped by project. Everything is directional.
Three workstreams carry the intelligence narrative (Active Main Hub, Proactive Yolanda, the Personalization Engine behind them). One is the polish that makes the whole product feel responsive (Navigation).
Users don't always know they can swipe, what direction, or where the bounds are. The fix is a polish pass and a bug catalogue.
- Swipe directional hints. Animated arrows show available cardinal directions.
- Haptic feedback on swipes. Adds the physicality the gesture currently lacks.
- eSIM Shop polish + shimmers. Style updates and shimmer loading states across the eSIM Shop.
- Hitbox and bounds indicators. Cues for where the swipe area starts and ends.
- Swipe bug fixes. Catalogued, prioritized, ship on their own cadence.
The hub today is crowded and static. The redesign declutters it and gives it a priority-stack slot that adapts to the user.
- Two zones. eSIM shop module on top, carrying the user's active promo as context. One dynamic module slot below.
- Priority stack. Acting on or dismissing a module pops the next one up.
- Declutter. "My eSIMs" and active subscriptions move to a profile-level utility page.
- Ambient, persistent. Modules wait until acted on. Not interruptive. The screen is simply like this now.
The eSIM shop module carries state.
The block stays prominent at the top of the hub, but its surface reflects the offer that's relevant right now. Default messaging when there's nothing to push, promo-led when there is.
Candidate modules in the stack.
The kinds of things that fill the slot. Trigger rules and persistence are PRD-level work.
Today Yolanda only acts when a user opens chat and types. Proactive Yolanda is the same voice reaching out, with a short prompt that offers one clear next step.
- Two states · Idle and Active. When Yolanda is reaching out or in conversation, the surface shifts to Active. Declutters the bottom of the screen and gives focus to whatever matters in that moment.
- Single nudge per session. Best for cold starts and exploration. Pointing toward an underused surface, or opening chat to help with a task.
- Speak from the situation, not the data. The system never reveals the signal that triggered the prompt.
- Tone varies by stakes. Warmer for trip insurance, breezy for a radio station.
- Skip is a signal, not an apology. One fading line, then it's gone for the session.
Search · Yolanda meets the map.
An updated search flow makes Yolanda and the map feel like one surface. Recommendations and search results show up as both result cards and interactive pins, so the user can move between shapes without losing context.
- One surface, two shapes. Recommendations render as cards and as interactive pins on the map. Yolanda search results render as map pins and as result cards similar to "recommended". The user moves between them without losing place.
- Named pins. Location pins on the map display the venue name, so the user can scan results without tapping.
- Improved location indicator. The user's location dot reflects movement direction and orientation, so the map reads correctly while walking.
- Back to Yolanda. Detail cards keep an explicit return path to the conversation that surfaced the result.
A shared personalization layer feeds both adaptive surfaces. Same signals about user state and behavior, interpreted two different ways.
Hub vs. Yolanda, at a glance.
| Hub | Yolanda | |
|---|---|---|
| Feel | Ambient | Interruptive |
| Cadence | Persistent stack | One per session |
| Lifespan | Sits until acted on | Appears, then gone |
| Best for | Offers, your stuff, saved & tracked things | Cold starts, exploration, conversational help |
V2 doesn't run alone. These are the parallel programs whose resourcing and store-submission cadence shape V2's actual landing dates. Movement on any of them ripples through.
Bancus
Static wallet in V2.2 (Jun 1). Dynamic wallet with eligibility in V2.3 (Jun 15). The major surface for the wallet flow this quarter.
Jazz App launch
Scope being defined. Shares engineering bench with V2.4 and V2.5. Anything that moves Jazz in or out moves V2's narrative releases with it.
GlobalYO web rework
Technical and branding rework. WordPress front-end retiring, replaced with a Next.js build aligned to the V2 brand. Independent of the app cadence but lands in the same review window.
KYC, eSIM only, new MVNO
x.x.1 brings KYC to all users (Jun 1). x.x.2 removes physical SIM (Jun 18). A third release at end of June integrates a new MVNO partner. Three store submissions inside one month.
Payment refactor
Payment refactoring and stabilization release. Foundational work that gates the personalization and dynamic-module work that follows.
Flights, insurance, accommodations
Insurance (V2.y) is the easier integration and lands first if the existing partner becomes responsive. Flights (V2.x) has been partially worked but is currently deprioritized, targeting September. Accommodations is a later phase, blocked by Duffel approval into their partner program.
No surprises here. Each one is a candidate for slipping the V2.4 or V2.5 window, or compressing scope inside it.
Engineering bench sharing
Jazz · Bancus · YoMobile · PaymentV2.4 and V2.5 share the same engineering capacity as JazzApp 1.0.0, the Bancus rollout, three YoMobile releases, and the payment refactor. Movement on any of those is movement on V2.
Rebrand-window scope
Per workstreamWhat's in for the window versus follow-on phase is the in-flight conversation. The dates assume an upper-end scope read.
Partner responsiveness
Insurance · DuffelInsurance is gated on the current partner becoming responsive. Accommodations is blocked on Duffel admitting us to their partner program. Both are outside our control.
Payment refactor dependency
V2.3.1 · pre-narrativePayment refactoring needs to ship before new feature development picks up downstream. A slip on V2.3.1 ripples into V2.4 and V2.5.
Data plumbing
Posthog tier 2 · tier 3Posthog tier 2 in V2.2 and tier 3 in V2.3 are the foundation for the engine's signal model. Late events limit how intelligent surfaces feel at launch.
V2.5 scope · the Active Main Hub
Concept stageThe Active Main Hub is at concept stage. The PRD follows agreement on the concept. A late concept lock pushes V2.5 or compresses what ships.