The Digital Twin Era needs more than great visuals.
Validation, development phases and data quality in automotive.
The layer that doesn’t show
There’s a statement worth putting at the center before going further: in the product development of the coming years, the differentiator won’t be the most advanced headset. It will be system quality.
We developed this reasoning in a previous piece — XR as infrastructure, not as event. This article starts from there and takes a more specific step: not what makes XR a valid system, but why data quality depends on something that comes before the hardware, before the software, before the research protocol.
It depends on the quality of what the participant has in front of them.
The paradox of immersive research
The promise of immersive research is to collect authentic reactions — behavioral, not just declared — on a product before it physically exists. It’s a solid promise, supported by evidence: reactions in a high-fidelity immersive environment are closer to real behavior than those produced by a questionnaire or a photo on a screen.
But that promise has a condition that is almost always underestimated: the environment must be faithful enough to produce genuine reactions.
A participant in front of a model with approximate proportions isn’t evaluating the vehicle. They’re mentally compensating for the representation’s imperfections. Their attention is partially occupied by what the model fails to communicate — not by what the product actually is.
In a structured research context, this isn’t an aesthetic problem. It’s a data validity problem.
Data collected on an imprecise asset isn’t data about the product. It’s data about the artifact. And a decision made on that data isn’t an informed decision — it’s a hypothesis with a veneer of method.
What changes across product development phases
Market research in automotive product development isn’t a single event. It’s a process that runs through distinct phases — and in each one, the questions change, the available assets change, and the type of useful data changes with them.
In the concept phase, when design directions are still open, the value of research lies in the ability to discriminate between alternatives: which architecture works for a given segment, which formal language is coherent with the expected positioning, where the product sits relative to market references. Assets at this stage are often partial, evolving. The method must be flexible enough to work on non-definitive stimuli while maintaining comparative rigor.
In structured design review, the product has a more defined form. Questions shift to perceived surface quality, zone coherence, material rendering. Here, asset fidelity becomes critical: an approximate model at this stage produces feedback that doesn’t transfer to the real product — generating false confidence or false alarms.
In the pre-launch phase, research changes nature: it’s no longer evaluating the design itself, but the market’s response to the product in its commercial configurations. Colorways, trim levels, perceived pricing, competitor comparison. Data here must be comparable across markets, structured by segment, readable by different teams — product planning, marketing, CMI.
Each phase has its own logic. Confusing them — using pre-launch methods in the concept phase, or collecting comparative data on still-incomplete assets — produces outputs that don’t support the decisions they’re meant to inform.
This awareness doesn’t come from technology. It comes from method. And method is built by knowing the process from the inside.
Why visual quality is not an aesthetic detail
In an immersive research system, digital asset quality is not an aesthetic variable. It’s a methodological one — with the same weight as the research protocol, the panel composition, the questionnaire structure.
A Digital Twin built to campaign standards — photorealistic material management, IBL-based lighting control, surface precision, real-time optimization without perceived quality loss — is the correct starting point for a research session. Not because it “looks better,” but because the reactions it generates are reactions to the product — not to a simplified version of it.
This matters increasingly as the development cycle advances. In the concept phase, a stylized model may be sufficient to discriminate between directions. In design review or pre-launch, surface rendering, bodywork response to light, and zone coherence across the vehicle are exactly what the participant is evaluating — consciously or not.
If that element isn’t in the scene, the participant can’t react to it. If it’s in the scene but imprecise, the participant reacts to the imprecision.
In both cases, the data collected isn’t data about the product.
Data as a measure of system quality
According to McKinsey’s August 2025 analysis on automotive product development, Chinese EV manufacturers have brought development cycles to roughly 24 months — half the 40–50 months of legacy OEMs. That compression isn’t just a logistical challenge. It’s a methodological one.
The windows available to collect structured market feedback narrow and move earlier. Validation tools need to adapt to a pace that the traditional clinic can’t sustain in early phases — for reasons of cost, logistics, timing, and asset confidentiality.
The answer isn’t faster research that sacrifices rigor. It’s building a system where methodological rigor and digital asset quality combine to produce reliable data within a timeframe compatible with real decisions.
A system like this requires competencies that rarely coexist in the same organization: the ability to build digital assets that meet research standards, and the ability to structure research protocols that know what to do with that data.
Who brings what
In2real has spent years building high-fidelity Digital Twins for complex products — not as standalone aesthetic output, but as visual infrastructure integrated into brands’ development and communication processes. The pipeline that guarantees photorealistic asset quality in real-time environments isn’t a capability that’s improvised: it’s the result of a path that runs through CGI, rendering, multi-hardware optimization, and deep understanding of materials and surfaces.
Intelligo brings 20+ years of automotive research: clinic methodology, field management, structured research protocols, interpretation of data that combines declared and behavioral signals. Knowing the product development cycle from the inside — understanding that the useful data in concept phase isn’t the same as in pre-launch, and structuring the protocol accordingly — is competence built in the field, not in a technology lab.
The combination of the two isn’t a partnership. It’s the answer to a structural problem: immersive research produces reliable data only when those who build the asset and those who build the research method speak the same language.
Follow us to dicover what
We’re building a tool that answers exactly these problems: a structured immersive research platform, designed to produce reliable data in the phases of the development cycle where traditional clinics don’t reach. It’s a path we’re following, one that in the coming weeks will move into field validation.
If the problem sounds familiar, it makes sense to follow what we’re building!
Reach out if you’d like to know more before the launch.
I believe great communication begins with listening. With a background in design and five years leading In2real’s communication efforts, I work with our team to help brands connect with their audiences through clear strategy, visual quality, and digital storytelling.