Forthcode Unveils ANiX AI Cabin Assistant

"AI That Talks" vs. "AI That Enables Action", Why Forthcode's Offline Cabin Assistant Is Actually a Bet Against How Most Airline Tech Gets Built

Forthcode has unveiled Forthcode ANiX, an AI intelligence layer built specifically to keep working when an aircraft's connectivity is limited or unavailable, a genuinely unusual design constraint in an industry where nearly every consumer-facing AI product assumes continuous cloud access as a given. The platform will get its first major cabin demonstration at ForsaTEK 2026 in Dubai, Emirates Group's annual innovation showcase, built through Forthcode's work inside the Emirates Aviation X Lab ecosystem and in collaboration with Thales and Forthcode Collins Aerospace, two of the aviation industry's most established inflight entertainment and connected-cabin technology suppliers. CEO Ajith Balakumar's own framing of the product's purpose is the sentence that actually matters here: "AI in aviation needs to do much more than answer questions. The real opportunity is connecting passenger intent with the systems, knowledge and people capable of acting on it."

That distinction, between AI that answers and AI that acts, is not marketing language. It describes a genuinely different, and genuinely harder, engineering problem than the chatbot wave that has already swept through airline customer service over the past two years.

Why "Offline-First" Is a Genuine Technical Constraint, Not a Feature Flourish

Most AI products built for consumer or enterprise use assume the device running them has reliable, continuous internet access, a reasonable assumption on the ground, but one that breaks down entirely inside an aircraft cabin cruising at 35,000 feet, where connectivity depends on satellite links that can be intermittent, bandwidth-constrained, or, on aircraft not yet retrofitted with modern systems, simply absent. This feed has covered exactly how significant that connectivity gap can be, Aer Lingus's Starlink rollout, where 85% of passengers connected and consumed up to 950GB of data on a single transatlantic flight, demonstrates what becomes possible once genuinely fast, reliable inflight connectivity exists. But Starlink retrofits take years to roll out across an entire fleet, and even airlines actively racing to install it, Aer Lingus itself targets full long-haul fleet coverage only by the end of 2026, will have thousands of flights in the interim running on older, patchier connectivity systems, or none at all.

ANiX's core engineering bet is that an aviation AI product built assuming that gap exists, rather than treating it as an edge case to patch around later, is more useful in practice than a cloud-dependent system that simply stops working the moment the satellite link degrades mid-flight. Building genuinely offline-capable AI, models and functions that can run locally on cabin hardware without a live connection back to a central server, is a materially harder technical problem than building a standard cloud-based chatbot, precisely because it requires compressing whatever AI capability is needed into something that can execute reliably on local, resource-constrained onboard systems rather than drawing on effectively unlimited cloud compute.

What a Passenger Asking About a Vegan Meal Actually Reveals About the Product's Ambition

The specific example Forthcode uses to illustrate ANiX's function, a passenger asking through the inflight entertainment system which onboard meal is vegan, or requesting a movie suitable for a child, sounds almost deliberately mundane. But the mundane framing is the point: these are exactly the kind of routine, low-stakes passenger queries that current-generation cabin chatbots already handle reasonably well using pre-programmed responses or basic language models. What makes ANiX's architecture different is what happens after the question gets answered. Forthcode's own description states that "where systems allow, the passenger intent can also be passed into crew applications, onboard inventory, catering or commerce systems", meaning a vegan meal query is not necessarily a dead-end interaction that produces a generated text answer and nothing else. It can, in principle, become a structured data point that flows into the actual operational systems crew and catering teams use, potentially informing real-time inventory awareness or triggering a crew notification, rather than existing only as a conversational exchange between a passenger and a screen.

That is the genuinely more ambitious and more technically demanding claim buried inside an otherwise ordinary-sounding customer service example, connecting a natural-language passenger request to backend operational systems that were, in most airlines' current infrastructure, never designed to receive input from a passenger-facing AI layer at all.

Why This Reflects a Broader Shift Forthcode's Own CEO Has Been Tracking Publicly

Balakumar's own recent industry commentary, published separately from the ANiX launch itself, makes the underlying thesis explicit: "For the past two years, much of the aviation industry's conversation about artificial intelligence has focused on customer-facing applications.

 The more significant shift now taking place is behind the scenes. AI is moving from answering questions to helping airlines make and execute decisions across increasingly complex operational environments." That framing positions ANiX not as an isolated product launch, but as Forthcode's specific attempt to build the connective infrastructure layer this broader industry shift requires, one that spans passenger-facing interaction, airline-approved knowledge content, and the operational backend systems (crew workflows, catering, inventory, commerce) that a purely customer-service-oriented chatbot never needed to touch.

Why Building This Requires Partners Who Already Own the Hardware Layer

Forthcode's collaboration with Thales and Forthcode Collins Aerospace is not incidental to this strategy, it is close to a prerequisite for it working at all. Both companies are established, deeply entrenched suppliers of inflight entertainment and connected-cabin systems across the global airline fleet; an AI layer attempting to connect passenger interaction to catering and crew systems cannot realistically bypass the hardware and software infrastructure those suppliers already control inside the cabin. Building ANiX in partnership with, rather than in competition against, the incumbent IFE and connected-cabin suppliers gives Forthcode a plausible path to actual deployment across real aircraft, a materially different position than a standalone AI startup trying to convince airlines to rip out and replace existing, expensive, already-certified cabin hardware just to accommodate a new software layer.

Why Emirates' Own Incubator Ecosystem Is the Testing Ground This Kind of Product Actually Needs

ForsaTEK itself, now running its fourth edition since Emirates Group launched it in 2023, exists specifically to showcase exactly this category of aviation-technology innovation, bringing together Emirates' Intelak and Aviation X Lab startup incubator programmes with established technology partners including, in past editions, Microsoft, Amadeus, Airbus and J.P. Morgan's blockchain unit. That track record matters for assessing ANiX's credibility: Forthcode is not launching this product into a vacuum or purely through its own marketing channel, but through an established platform Emirates has spent years building specifically to validate genuinely novel aviation technology concepts in front of the airline industry's own decision-makers, before those concepts have necessarily proven themselves at commercial scale.

Being demonstrated through Aviation X Lab specifically, rather than as an independent product launch, also signals Forthcode has likely already been developing and refining ANiX with direct input from Emirates' own technical and operational teams over some period, giving the eventual cabin demonstration a level of real-world grounding that a purely externally developed product debuting cold at an industry conference would lack.

What Would Actually Prove This Works, Beyond a Trade Show Demo

The genuinely hard test for ANiX is not whether it can convincingly answer a vegan-meal question during a controlled ForsaTEK demonstration, that is a solvable problem current-generation AI already handles reasonably well. It is whether the claimed backend integration, passenger intent flowing into actual crew, catering and inventory systems, can function reliably across the genuinely messy, varied, often legacy IT infrastructure that different airlines run behind their inflight entertainment systems, and whether that integration can be certified and deployed at the safety and reliability standards commercial aviation requires for anything touching operational systems, not just passenger entertainment. Forthcode's own product description elsewhere describes its AI agents as running "on your knowledge" and following "your rules" without data mining or vendor lock-in, language aimed directly at addressing the data governance and airline-control concerns that would be the first serious objection any airline's IT and compliance team would raise before allowing a third-party AI layer anywhere near crew workflows or catering inventory systems.

Whether ANiX moves from a ForsaTEK cabin demonstration into genuine, revenue-generating airline deployments will depend less on the underlying AI capability, which the offline-first architecture and the Thales/Collins Aerospace partnerships suggest is technically serious, and more on whether Forthcode can convince individual airlines' operational and IT teams that connecting a passenger-facing AI assistant to real backend systems is worth the integration complexity and governance risk that connection inherently introduces. That is a considerably harder sale than shipping another cabin chatbot, which is precisely why Balakumar's own framing draws such a deliberate line between the two.