Service · AI development

Developing home automation apps for smart homes and smart buildings.

One custom app to control lighting, climate, security, energy and presence, for homes, elderly care, hotel rooms and commercial buildings. It works across Matter, Thread, Zigbee, Z-Wave and KNX, with an AI layer for forecasting and anomaly detection.

Matter & ThreadZigbee / Z-WaveKNX & BACnetAI layer

Arrange an introductory call See our work

A home automation app is not a collection of separate brand apps.

The reality in a home or building is almost always a mix: a Hue lamp here, a Tado thermostat there, a Niko switch in the wall, KNX wiring in the ceiling joists, a few Aqara sensors in the hallway and a commercial BACnet controller on the air handling unit. Four or five apps to operate all of that is not a good experience, it is a nuisance. A well-built home automation app abstracts that underlying layer away and gives the user one coherent interface, whether that user is a resident, a care worker, a hotel guest, an installer or a facilities manager.

We develop home automation apps for both consumer contexts and commercial smart buildings. For care institutions that use home automation for getting-up, fall and shower detection. For housing associations delivering new builds and renovations with a digital package. For hotel chains that want to save energy through presence detection and room presets. For developers who use smart home as a selling point at handover. And for installers who want to give their clients one clear app instead of a stack of manuals.

This page is part of our wider AI services: home automation without AI is a remote control, home automation with AI is a building that adjusts itself based on patterns, weather forecasts, energy prices and behaviour. That AI layer is exactly what distinguishes the product of an old-fashioned home controller from a truly smart system.

What makes our work different from a screen for a single protocol: we build multi-protocol from day one, integrate with existing care, property and energy systems, and keep hardware freedom open for the client. No vendor lock-in where the resident or client is tied to one lighting brand. The device portfolio can change in future without a problem: our app responds with a driver update, not a new project.

Three types of home automation app we build.

The choice depends on who the end user is, which protocols are already in the walls, and how central AI should be to the control. In the first conversation we decide together which variant fits.

Consumer app · smart home for residents

Smart home app for residents or tenants

One app for the home: lighting, climate, blinds, security, doorbell, smart meter, EV charger and presence simulation. We build on Matter and Thread so that Apple Home, Google Home and Alexa work natively, with Zigbee and Z-Wave as a fallback for existing devices. For housing associations we deliver the app as a white-label product so that the tenant associates it with the association and not with a hardware brand, a logical extension of the work we do around housing association software.

Matter / ThreadZigbee & Z-WaveApple Home / Google HomeWhite-label

Care automation · independent living for older adults

Care automation for sheltered housing

Smart home technology for independent living, with care working quietly in the background: motion sensors in the bedroom and hallway that detect getting up and restlessness, shower monitoring with an absence alert, fall detection via radar or floor sensors, fridge door and microwave cycle tracking for routine analysis, and alerts sent to the control room or family carers. We integrate with the care software the organisation already uses, so alerts land in the same record timeline. More on that context is on our page about building custom care software. The AI layer spots pattern deviations before they become an incident.

Fall detectionRoutine AIFamily carer portalNEN 7510

Smart building · hotel, office or care complex

Building management app for commercial properties

A central app for smart offices, hotel rooms or care facilities running on KNX or BACnet. Hotel guests can control climate and ambience through their own app or an in-room tablet, energy is automatically reduced when a room is empty, and operations see in real time which meeting rooms are occupied. Under the hood, we integrate with the building management systems already in place and with the installer's ventilation and cooling, without building yet another silo on top.

KNX / BACnetOccupancy AIHotel presetsEnergy monitoring

What you get at the end.

A production-ready home automation app with its cloud layer, integrations with your existing systems, and everything around it so you can manage and expand it yourselves, including new protocols or hardware suppliers.

For installers who want to support their engineers in the field, we often pair the smart home app with a separate service app. See also our page on field service app development. There, the engineer sees in real time which device is running, what firmware level it has and which room the resident reported as a problem.

The app itself, iOS and AndroidCross-platform with React Native or Flutter, or native where widgets, Apple Watch control or complex sensor integration call for it. Publication on the App Store and Google Play included.
Multi-protocol cloud layerA single backend that communicates with Matter, Thread, Zigbee2MQTT, Z-Wave JS, KNX-IP and BACnet/IP. A local gateway in the home or building enables offline control when the internet goes down.
Home Assistant or your own brokerFor demanding integrations, Home Assistant often runs under the hood as a universal broker. We manage that integration layer, so you only see one coherent API.
AI layer for forecasting and pattern detectionModels for energy forecasting based on weather predictions, occupancy learning rules, anomaly detection for burglary or water leaks, and personalised scenes that adapt to the resident's behaviour.
Voice integrationAlexa, Google Assistant and Siri via official certification, plus optionally a locally run LLM for privacy-sensitive voice control in care environments.
Smart meter and EV chargingIntegration with the P1 port and with OCPP charging points, so the app shows energy consumption and generation and adjusts charging to solar surplus or dynamic tariffs.
Operator management portalFor the care organisation, hotel chain or housing corporation: user management, alarm routing, firmware rollout across the whole fleet, and reporting on energy, occupancy or incidents.
Codebase, runbook and ongoing development contract (optional)Full source code, architecture overview, an incident runbook and ongoing capacity for new drivers, model updates and feature expansion.

When a custom home automation app is the right choice.

Four patterns in which we guide clients. If you recognise one of them, we are happy to talk through what the right answer is for your situation.

Multi-protocol

You have multiple systems at home

Off-the-shelf apps from manufacturers only support one ecosystem. As soon as you want to combine Hue, Tado, a KNX installation, a smart meter and an EV charger, you need one app that can communicate across all of those worlds. That is exactly what our multi-protocol architecture is built for.

White label

You want to deliver under your own brand

A housing association, property developer or installer wants residents or customers to associate the app with their brand, not with a hardware manufacturer. On handover of a new-build or renovation project, that means an app in your own visual identity, with your own support routing.

Care context

You combine home automation with care

Care-focused home automation goes beyond comfort: activity pattern detection, fall alarms, shower monitoring, and alerts to the emergency control room or informal carers. This falls under NEN 7510, GDPR and, for medical decisions, the AI Act's high-risk classification. A standard app does not meet those requirements.

Scale

You manage dozens to thousands of units

You can still configure a single home by hand. For a complex of 80 care homes, a hotel with 200 rooms or a housing association portfolio of thousands of addresses, fleet management of firmware, drivers, users and alarms becomes indispensable. That is built into our back end.

Tech stack and key decisions.

We are not ideological about one protocol or platform. The right choice depends on which hardware is already in place, how sensitive the data is and how quickly you want to scale. Below are the choices we often make, and why.

For the broader energy layer (dynamic tariffs, smart meter data, generation and control), we often combine the home automation app with patterns from custom energy software.

Mobile: React Native, Flutter or nativeReact Native for a single codebase with fast iteration, Flutter for pixel-perfect UI and strong animation performance, and native Swift and Kotlin where Apple Home extensions, widgets or Apple Watch control are decisive.
Protocols: Matter, Thread, Zigbee, Z-Wave, KNX, BACnetMatter and Thread are the future for consumer devices and are backed by all the major platforms. Zigbee and Z-Wave will remain relevant for years for existing installations. KNX and BACnet are standard in commercial and professional buildings, and we integrate them via IP gateways.
Local broker: Home Assistant or a custom buildFor projects with many sensor types and complex automation, we run Home Assistant as a local broker. It offers a vast driver library and local fallback when the internet goes down. For simpler projects, we build the broker ourselves, with fewer dependencies and less maintenance.
Cloud backend: Node.js, Python or GoPython for AI models and data processing, Node.js for real-time WebSockets and MQTT bridges, Go where throughput performance matters for large fleets. A scalable event bus in between, such as RabbitMQ or NATS, keeps protocol translation decoupled from business logic.
AI models: forecasting, anomaly detection, voiceTime-series forecasting (Prophet, LightGBM) for energy prediction, isolation forest or autoencoder for anomaly detection of intrusion and leaks, and optionally a local LLM (for example a quantised Mistral or Llama model) for offline voice commands in care settings.
Voice: Alexa, Google Assistant, Siri and local LLMThe big three through official certification for broad accessibility. In addition, a locally running LLM for situations where speech may not be sent to the cloud, for example in a care home where residents' utterances are privacy-sensitive.
Smart meter and EV: P1, DSMR, OCPPReading the P1 port via a hardware dongle or smart meter API, integrating with OCPP charging points to steer by solar surplus and dynamic tariffs based on EPEX or EnergyZero wholesale prices.
Key decisionsLocal or cloud-first, one gateway per building or a shared cloud broker, which protocol serves as the primary, what to do during an internet outage, how you roll out firmware across thousands of devices, and how the GDPR and the EU AI Act apply to resident behaviour data. We make these trade-offs together, early on.

Compliance for a home automation app — what we take care of.

Home automation touches more regulation than a standard consumer app, because behavioural data is highly sensitive and certain care applications fall under strict requirements. We map out early in the process what applies.

GDPR

Behavioural patterns are sensitive data

Presence, movement and daily-routine data reveal a great deal about someone's life and, under certain definitions, may be treated as special category personal data. Data minimisation, purpose limitation, short retention, processor agreements and a DPIA in a care context are standard parts of the process.

AI Act

Care-oriented home automation can be high-risk

When the AI layer makes decisions that affect a resident's health, for example raising an alarm on a suspected fall or automatically escalating a deviation in daily routine, this may qualify as high-risk under the EU AI Act. That means more demanding documentation, evaluation and demonstrable human oversight.

NEN 7510 & ISO 27001

Security in care settings

For care-oriented home automation, NEN 7510 is the standard for information security. We work with data minimisation, encryption in transit and at rest, an audit trail for sensitive actions, and clear role separation between care staff, informal carers and residents.

CE marking and AVCT

Hardware and care installations

Hardware components carry their own CE marking, and we integrate with approved devices. Specific smart home installations in healthcare sometimes fall under a public tendering obligation (AVCT) or an additional quality framework. We work with installers who already hold that certification.

How a home automation app project runs with us.

Introduction and context mapping

A conversation to understand which buildings, which users and which protocols are involved. Often this already includes an initial assessment of consumer product versus care context and smart building, plus a check on GDPR and EU AI Act implications.

Workshop with installer and users

We spend a morning with your team, an installer who knows the hardware, and a few end users, whether residents, care staff or facility managers. At the end you have a defined scope, a first screen flow and a short list of technical decisions to investigate first.

Prototype with real hardware

We build a tangible prototype flow on a test setup using the hardware you actually plan to use. Not a mock-up, but a working slice in which lighting, climate and one AI feature already run. This is where we validate the protocol choices before building at scale.

Building in sprints

A working build every two weeks, tested on iOS, Android and the physical hardware. In parallel we work on the AI models, the management portal and protocol extensions. You test along the way, often with a small group of end users in a pilot setting.

Pilot, rollout and ongoing development

First a pilot in a handful of homes or one hotel wing. Then a phased rollout across the full portfolio, with over-the-air firmware updates and monitoring from the management portal. Ongoing development on a fixed cadence for new drivers, model updates and feature requests from the field.

Frequently asked questions.

What clients usually want to know before we start.

What is the difference between a home automation app and a regular smart home app?

A manufacturer's smart home app (Hue, Tado, Niko) controls a single ecosystem. A custom smart home app is protocol-agnostic and integrates across Matter, Zigbee, Z-Wave, KNX and BACnet, with one coherent interface for the end user and one management portal for the operator. The difference lies not in the UI but in the integrations running under the hood.

Which protocols do you support?

Matter and Thread for the latest consumer devices, Zigbee and Z-Wave for existing installations, KNX for professional residential and commercial building work, BACnet for commercial building management systems, MQTT for custom integrations, and HTTP/REST and webhooks for cloud-to-cloud integration with brands that don't offer an open protocol.

What does the AI layer actually do?

A few examples from real projects: energy forecasting based on weather forecasts and historical consumption, presence detection that learns when a room is actually in use so that climate and lighting adjust accordingly, anomaly detection for unusual water consumption (leaks) or intrusion-like movement patterns, and, in a care context, living-pattern models that flag, without a camera, that a resident has been in the shower unusually long or has not got out of bed.

Does the app still work if the internet goes down?

Yes, for the core functions. A local gateway in the home or building runs critical automations, basic control and alerting even when the internet goes down. Cloud features such as voice assistants, AI forecasting and remote access do require a connection, but the lights still work, the thermostat still responds and the alarm still sounds without internet.

What about the privacy of behavioural data?

Presence and living-pattern data is sensitive, and in a care context we treat it as special category personal data. By default we apply data minimisation (only what is necessary), short retention, EU data residency, a data processing agreement with you and, for care home automation, a DPIA before rollout. For the most sensitive steps, processing runs locally rather than in the cloud.

Can we supply it white-label under our own brand?

Yes. Many of our clients, including housing associations, hotel chains, installers and builders, deliver the app under their own branding. Your own icons, colours, name in the App Store and Google Play, and your own support routing. We share the technical foundation behind the scenes, and the front office is entirely yours.

Which compliance requirements apply to care home automation?

GDPR for behavioural data, NEN 7510 for information security in care, and the EU AI Act where the AI layer makes decisions that affect a resident's health or safety, which may fall under high-risk. Specific care installations sometimes require a public tender (AVCT). We map this out at the start and build in the right documentation and human oversight. For a broader care software framework, see our page on building care software.

Do you work together with our installer?

Almost always. The installer knows the hardware, the wiring and the building; we know the app, the cloud and the AI layer. We work with a shared architecture drawing and a clear division of roles: the installer delivers the physical layer and the cabling, we deliver the digital layer. For the service phase we often add a dedicated field service app so that engineers in the field have direct access to device status and history.

What determines the cost of a home automation app project?

Mainly four things: the number of protocols we need to integrate, the complexity of the management portal (a single building or thousands of units across a fleet), whether an AI layer is added on top and how deep it goes, and the compliance context (consumer is lighter than care). We work with sprint budgets and, after the first conversation, give you an order-of-magnitude estimate before you make any commitments.

Can you also manage and further develop the app?

Yes, and for home automation apps that is almost always sensible. New protocols keep emerging (Matter adds new device types with each release), AI models improve, and your hardware portfolio will change. An ongoing development contract with fixed sprint capacity keeps the app moving with the market, and gives you a fixed point of contact when a new driver or integration is needed.

Fabian van Dijk Business Developer · Appfront

Questions about your domotics context, a second opinion on an ongoing quote, or simply help thinking through the scope? Email fabian.vandijk@appfront.nl or book an introductory call via /contact.

Talk to us about your domotics project.

A free, no-obligation half-hour introductory call. We'll listen to which buildings or homes you have in mind, help you weigh protocol choices, the AI layer and compliance, and point you in a useful direction — even if the right next step isn't a full build, but first a prototype on a single building or a short validation sprint. Whether you're an installer who wants to deliver one app to your clients, a housing association delivering new-build homes, or a care organisation supporting independent living, we're happy to have the conversation before everything is set in stone.

Book an introductory call Or see more of our work