The device is the easy part. What decides whether a connected product works is everything behind it: how firmware gets updated in the field, what happens to telemetry when the network drops, and whether anyone can act on the data once it finally arrives.
The software layer of a connected product, from the device credential up to the dashboard.
Telemetry ingested, stored at a resolution you can actually afford to keep, and turned into dashboards and alerts people act on. A fleet producing data nobody reads is a cost centre, not an asset.
Provisioning, identity, configuration and over-the-air updates for devices you cannot physically reach. Getting this wrong is how one firmware bug becomes a van journey to every site you have.
Companion apps and remote monitoring for devices people wear, where battery life, sync behaviour and the handling of health-adjacent data set the constraints before any feature does.
Named up front, because avoiding them is most of the value.
A fleet that cannot be updated over the air turns every defect into a field visit. This has to be designed in before the first unit ships, not retrofitted after the first recall.
Devices lose signal. Without local buffering and replay that cannot duplicate, an outage becomes permanently lost readings and double-counted records.
Full-resolution telemetry retained indefinitely is the fastest route to a storage bill nobody approved. Retention and downsampling belong in the design, not in a later cost review.
Share one credential across a fleet and a single extracted device compromises all of them. Per-device identity and rotation are not optional extras.
Every phase is run by senior engineers and ends in something you can see.
What the device must measure, how often, on what power budget and over which network — the constraints that quietly determine everything downstream.
Message schemas, sampling rates, retention and downsampling agreed against both the questions you want answered and the bill you are willing to pay.
Protocols, provisioning, per-device credentials and rotation, plus exactly how a device behaves when it cannot reach the platform at all.
Ingestion, storage, processing, rules and alerting, sized for the fleet you expect rather than the twelve-device pilot in front of you.
Operator dashboards, admin tooling and companion apps built around the decisions people actually make, not around every metric available.
Staged deployment, monitoring, the over-the-air update pipeline, and a runbook for the first time something fails at three in the morning.
Chosen per project, and for what your team can maintain afterwards.
No — we are a software company. Firmware integration, connectivity, the platform, dashboards and companion apps are ours; we work alongside your hardware team or manufacturer. That division of responsibility is agreed at the start so nothing falls between the two of us.
It buffers locally and replays when the link returns, with messages designed so that a replay cannot create duplicates. Assuming a reliable network is the most common and most expensive mistake in a connected product.
Yes, and we would argue it is the single most important thing to design in early. Staged rollout, rollback and per-device update reporting are what stop a defect turning into a visit to every installation.
Per-device identity rather than a shared fleet credential, encrypted transport, key rotation, and least-privilege access on the platform side. One extracted device should compromise exactly one device.
Whichever you already run, in most cases. AWS, Azure and Google Cloud all have adequate IoT primitives, and the deciding factors are usually your existing commitments, data residency and where your team’s skills already sit — not the feature matrix.
Often, depending on what the firmware exposes and whether it can be updated at all. It starts with an assessment of the existing fleet and its protocol, after which you get a written view of what can be brought in and what needs replacing.
iOS and Android products built to pass store review, launch on schedule and survive the annual OS churn.
Custom platforms, portals and SaaS — including replacing a legacy system while the business running on it stays up.
Senior engineers directing autonomous agents that plan, write and test in parallel — with a named human accountable for every release.