AI, Data & Emerging Technology · Emerging technology products
Turning new technology into a working product
We help companies test and build products on emerging technology such as spatial computing, on-device AI, voice interfaces and connected hardware, using short research spikes to separate what is feasible today from what is still hype.
- Feasibility spikes
- On-device AI
- Voice and multimodal interfaces
Overview
Separating promising technology from expensive detours
New technologies arrive with strong claims and immature tooling. On-device AI, spatial computing, voice and multimodal interfaces, and new connectivity standards can all open real product opportunities, but early SDKs change quickly, documentation is thin and performance on real hardware often differs from the launch videos. Vendor roadmaps can also shift, leaving early adopters to work around missing features or sudden deprecations.
The main decision is how much to commit before uncertainty is reduced. We favor short, time-boxed research spikes that answer one risky question each, such as whether a model runs fast enough on a phone or whether hand tracking is precise enough for a task, before any full build. That keeps spending tied to evidence and makes stopping or changing direction an acceptable outcome.
Good emerging-tech work produces findings you can act on, even when the answer is not yet. Each experiment ends with written conclusions, measured data and code you keep, so the knowledge stays with your company. Even a decision to wait comes with clear conditions for when it is worth looking again.
Who it’s for
Built for teams like yours
- 01
Product leaders under pressure
Executives and product managers asked how their company should respond to a new technology, who need an informed, hands-on view rather than another trend report.
- 02
Startups building on new platforms
Founders whose product depends on a recently released device, model or SDK and who need engineers comfortable working with incomplete documentation and frequent changes. We document every workaround we rely on.
- 03
Internal innovation groups
Corporate R&D and innovation teams that want extra hands for experiments, an outside view on feasibility, or help turning a successful prototype into a real product plan.
Why it matters
Find out early what is real
New platforms come with impressive demos and thin documentation. The risk is spending months on something the hardware, SDK or market cannot support yet. We reduce that risk with time-boxed spikes that answer one hard question at a time, such as tracking accuracy, battery life or on-device model speed, before committing to a full build.
When the evidence says wait, we will tell you plainly, and suggest a simpler version you can ship now while the platform matures, the tooling improves and the risks shrink.
Every engagement includes
- Opportunity framingthe user problem and the role the technology plays in solving it.
- Risk rankingtechnical, market and platform risks listed and ordered for testing.
- Research spikesworking experiments with written findings and a go, adjust or stop call.
- Prototypea functional version real users can try to validate the concept.
- Build planarchitecture, scope and phased roadmap for the first production release.
- Handovercode, findings and accounts delivered so your team keeps the knowledge.
Features
How we approach new tech
- 01
Feasibility spikes
One- to three-week experiments that test the riskiest technical assumption with working code.
- 02
On-device AI
Models running locally with Core ML, TensorFlow Lite or ONNX for privacy and offline use.
- 03
Voice and multimodal interfaces
Speech, vision and text inputs combined into products that respond naturally to people.
- 04
Spatial computing apps
visionOS and Android XR applications designed around physical space, eye gaze and hand gesture input.
- 05
Hardware-connected products
Apps and services that pair with sensors, wearables or custom devices over BLE and Wi-Fi.
- 06
Technology radar
A written view of which tools are ready, which to pilot and which to watch for your roadmap.
In practice
Kinds of exploration we run
On-device model feasibility
We test whether a language, vision or speech model can run on your target phones, laptops or edge hardware at acceptable speed, battery use and quality, compared with a cloud version.
Voice-first workflow prototype
A working prototype lets users complete a real task by speaking, such as logging field notes or placing a reorder, to test accuracy, speed and whether people actually prefer it.
Spatial computing app concept
We build a focused app for a new headset platform to learn what its input, display and passthrough capabilities make possible for your use case, and where the limits are.
New platform integration test
We connect your product to a newly released API, protocol or device ecosystem in a contained experiment, checking reliability, gaps and maintenance burden before you commit. Findings include an estimate of ongoing maintenance effort.
Process
How we work
- 1
Assumption mapping
We list what has to be true for the idea to work, technically, for users and for the business, and rank the assumptions by how uncertain and how critical each one is.
- 2
Time-boxed spikes
The riskiest assumptions are tested first through small builds lasting days to a few weeks each, with a clear question, success criteria and written findings at the end. Spikes are kept small on purpose.
- 3
User-facing prototype
Once technical risk is lower, a functional prototype is put in front of target users to test whether the experience is valuable and usable, not just possible. Observations are recorded during each session.
- 4
Decision review
We present findings, open risks and options, including stopping, waiting for the technology to mature, or proceeding, with a recommendation and its reasoning. The decision stays yours, informed by evidence gathered so far.
- 5
Product roadmap
If proceeding, we write a phased plan covering architecture, team needs and platform risks to watch, ready for our team, yours or a combination to execute. Platform risks include named triggers for revisiting plans.
Deliverables
What you receive
- Ranked assumption and risk map
- Research spike code and findings
- Benchmark data on target hardware
- Functional user-facing prototype
- User testing notes and observations
- Go or no-go recommendation memo
- Phased architecture and product roadmap
Tools & methods
On-device & AI
- Core ML
- TensorFlow Lite
- ONNX Runtime
- llama.cpp
- MediaPipe
- Whisper
Platforms
- visionOS
- Android XR
- Meta Quest
- Matter
- WebGPU
Methods
- assumption mapping
- research spikes
- benchmarking
- usability testing
- technology radar
FAQ
Frequently asked questions
Anything else about Emerging technology products? Ask us directly.
We test it against your specific requirements rather than vendor claims. A short spike measures what matters, such as accuracy, speed, battery use or device support, and ends with a written recommendation. That gives you evidence to decide whether to build now, adjust the idea or wait.
Let’s work together
Have a project in mind?
Book a strategy call and we’ll show you exactly how to turn your goals into a system that generates consistent results.