How we work
Start with the answer you need.
Then establish what it takes to deliver it reliably.
Industrial problems should not have to fit the capabilities of an existing sensor or monitoring product.
AxioIX begins with the operational answer or decision an organisation needs. We determine what physical evidence, operating context and analytical methods can support that answer, then investigate and validate a practical solution.
The objective is not simply to complete a demonstration. It is to establish what works, what value it creates and how the capability can operate dependably.

Our approach
Start with the answer. Establish the evidence. Prove the value. Engineer for repeatability.
Not every application needs new hardware, a trained AI model or a new communications system. The answer required decides what is needed.
02
Establish the evidence
What evidence would support that answer, and where it can come from.
03
Prove the value
Whether it works, whether it is worth doing, and whether it stands up to verification.
04
Engineer for repeatability
How to deliver the proven answer dependably, using the simplest suitable route.
Define the answer
What do you actually need to know or decide?
The first task is to establish the operational question. That means understanding:
- What the organisation needs to know, diagnose or decide
- Which equipment, process or service is affected
- Who needs the answer and what action they can take
- How quickly the answer must be available
- What confidence or accuracy is required
- What happens if the answer is wrong or arrives too late
- What operational or commercial value would justify the capability
Illustrative example
From a request to the real question
A customer may initially ask for vibration monitoring on pumps and motors.
The more useful question might be: can these assets continue supporting the cooling service reliably, what failure conditions matter, and how much warning do we need to act?
That changes which measurements are required, the interpretation needed and the eventual solution.
An anonymised illustration, not a described customer deployment.
Establish the evidence
What evidence would support that answer?
The required answer determines the evidence architecture: what must be measured or collected, in what context, at what fidelity, and how the result will be checked.
- Existing operational systems
- Process measurements, equipment operating states, control-system data, maintenance records, historical incidents and existing customer information.
- Physical measurements
- Vibration, movement, temperature, pressure, electrical current, flow or other relevant physical variables.
- Operating context
- Equipment configuration, duty cycle, operating conditions, material behaviour, environmental conditions and relationships with supporting processes.
- Independent outcomes
- Inspections, laboratory results, maintenance findings, production quality, process performance and subsequently observed conditions.
Conceptual illustration.
Evidence fidelity. The measurement detail needed to discover or understand a problem may be greater than the information needed for routine operation. High-frequency vibration measurements may be essential during diagnosis or model development, while a validated monitoring solution may need only selected features or event indicators. Richer evidence must still remain available, or be captured again, when exceptions or consequential events need investigation. Reduced routine sampling cannot always reproduce the original evidence.
Investigate and interpret
Use the engineering and analytical methods the problem requires.
The method is selected according to the evidence and the answer required. Depending on the problem, investigation may use:
- Existing customer instruments and information
- Suitable off-the-shelf or temporary sensing equipment
- AxioIX sensing products
- Laboratory and representative equipment testing
- Engineering calculations
- Signal processing and statistical analysis
- Physical modelling and simulation
- AI or machine learning where justified
Physical modelling
Some physical conditions cannot be observed directly during operation. Measured evidence may then be combined with an engineering model or simulation, with model output kept distinct from measured evidence.
Mill Intelligence is an example of this approach.
AI: a model is not yet a service
An AI model may be trained on suitable evidence to recognise or estimate a particular condition. The operating solution must then establish what measurements are required, how the model is configured, what context it receives, how its outputs are interpreted and how results are checked.
Training a model does not by itself authorise operational action, and not every application needs a newly trained model.
POC and POV
What must the proof establish?
Proof of Concept and Proof of Value answer two different questions. They are not necessarily separate, consecutive projects: they can be combined, or their order and scope adapted to the application.
Conceptual illustration.
Proof of Concept — POC
Can the proposed capability work?
A POC tests technical and analytical feasibility. It may establish:
- Whether the physical evidence can be captured
- Whether the measurement approach is suitable
- Whether information can be processed and communicated
- Whether an analytical method produces the required answer
- Whether the answer can be independently checked
- Whether installation, power, connectivity and environmental requirements are practical
A POC should have a defined scope, test conditions, acceptance criteria and a decision about the next step. A successful POC does not by itself establish commercial viability or production readiness.
Proof of Value — POV
Does the capability create sufficient operational value?
A POV evaluates whether the capability justifies investment or adoption. It considers:
- The value of the improved answer
- Whether the organisation can act on it
- The potential impact on reliability, process performance, quality or operating cost
- The effort and cost of deployment
- Ongoing operating and support costs
- Whether the result is sufficiently dependable
- The evidence needed to justify expansion
Where practical, results are compared with an appropriate baseline and independently observed outcomes.
Constrained environments
When experimental hardware cannot be installed.
Some industrial environments do not allow uncertified or experimental hardware to be installed. In those cases, AxioIX may investigate technical feasibility and potential value in other ways, and use the results to inform decisions about subsequent product engineering.
Route
Existing customer data
Operating, maintenance, process and inspection records that help establish the problem and its potential value.
Route
Approved instrumentation
Instruments already approved for the environment, where they can capture the relevant evidence.
Route
Laboratory testing
Controlled tests that establish whether the relevant signal or behaviour can be observed.
Route
Representative equipment
Comparable equipment or conditions outside the constrained environment.
This work can support decisions about subsequent product engineering. It does not replace the field validation, certification or qualification the final product will require.
Prove the answer
A useful answer must stand up to verification.
Producing a measurement, dashboard, model output or simulation does not by itself establish that the answer is correct. AxioIX tests results against independent evidence: physical inspection, repeated measurements, maintenance findings, laboratory tests, process conditions, quality results and subsequent operating performance.
The proof should test normal and abnormal conditions and, where relevant, examine false alarms, missed conditions, uncertainty, data availability and the effects of changing operating conditions.
At the end of the proof, the organisation should be able to decide whether to:
- Proceed
- Extend testing
- Revise the solution
- Use an alternative or simpler method
- Stop
Conceptual illustration. An analytical recommendation informs a decision; it does not by itself authorise action. Any automated intervention would require separately designed, authorised and validated control integration.
Engineer for repeatability
Once the answer is proven, choose how to deliver it repeatedly.
A successful POC does not automatically justify building a new sensor product. The delivery route depends on technical, operational and economic requirements.
Route A
Existing technology
Suitable existing instruments, customer systems, software or operational processes, where these can deliver the validated answer dependably.
Route B
Integrated solution
Established instruments combined with data processing, analytical methods, software and the relevant workflows.
Route C
Application-specific product
Where existing solutions are unsuitable or a dedicated product is commercially justified: the necessary sensing, communications, embedded-processing and integration components.
Conceptual illustration.
Reusable foundations
USP and rfFlux, from the first proof onwards
The Universal Sensor Platform is AxioIX’s continuously evolving architecture for sensing and embedded-product engineering. rfFlux is AxioIX’s already-developed, adaptable communications foundation.
Where new sensing or communications capability is needed, they are used during the initial POC, not only after it. Each application reuses established components and adds the hardware or software it requires; suitable new components become reusable for future applications.
The established components behind them are described on Technology. Reuse does not remove the need for integration testing, environmental qualification, cybersecurity assurance or relevant certification.
Economic considerations
A repeatable product should be justified against:
- Expected operating value
- Potential scale and product demand
- Hardware, software and installation costs
- Maintenance and support requirements
- Lifecycle obligations
- Environmental and regulatory constraints
- Opportunities for component reuse
Deploy, verify and expand
Make the capability dependable in real operations.
A successful demonstration must still become a reliable operating capability. That requires attention to:
- Installation and commissioning
- Measurement quality and calibration
- Data availability and communications
- Security and system access
- Operational responsibility
- Human authorisation and escalation
- Maintenance and support
- Configuration and software lifecycle
- Ongoing verification against real outcomes
Accountability stays with people. Responsibility for operational action remains with the appropriate people and organisations unless specific system authority has been established.
After deployment, evidence continues to inform adjustments, corrections and future development.
Where justified, a solution may be extended to additional equipment, supporting plant, larger asset populations or further sites. Each extension is a decision in its own right, not an automatic enterprise-wide roll-out.
Have an operating challenge to prove?
Tell us what you need to know or decide. We can help define the question, identify the evidence already available and scope a suitable first POC or POV.
That first engagement should leave you with clarity on the answer required, the investigation or proof needed, the acceptance criteria and the next decision.
