Skip to main content

Customer-Owned Data. Customer-Owned AI.

Data privacy discussions have largely focused on protecting individuals. This remains essential, but it is no longer sufficient. As analytics platforms and AI systems mature, a new question has become unavoidable: who owns the intelligence derived from business data?

The question is often framed as benchmarking. If enough data exists across similar businesses, locations, or industries, should it be used to identify trends, best practices, or performance baselines? From a technical standpoint, this is entirely possible.

From a principled standpoint, it is not acceptable.

Benchmarking requires reuse. Reuse requires that one customer’s operational data, directly or indirectly, improves the outcomes of another. Even when raw data is anonymised and never shared, shared model training and global optimisation still transfer value.

This is not a privacy issue in the traditional sense. No individuals are identified. No personal data is exposed. Yet competitive intelligence is created and redistributed.

Modern analytics systems capture behavioural patterns, spatial dynamics, and operational performance over time. Accumulated at scale, this data becomes a precise representation of how a business operates in a specific context. That intelligence is proprietary by nature.

The core principle is therefore simple:

Customer-owned data must be used only to build models for that customer.
Any intelligence derived from that data must remain exclusive to that customer.

This principle applies to raw data, derived features, behavioural abstractions, and trained AI models. Model training itself is a form of data use. If models are shared, value is shared.

This position does not prevent the use of external or public datasets. Weather, economic indicators, demographic statistics, transport schedules, and geographic data describe the environment. They do not describe a customer’s operations.

Public context may be shared. Proprietary behaviour may not.

This approach goes beyond regulatory compliance. Laws define minimum obligations. Trust requires clearer boundaries.

As AI systems scale, there is strong incentive to aggregate and generalise. Resisting that incentive is a deliberate design choice. It prioritises customer autonomy over platform leverage.

Analytics should make customers more competitive, not more transparent to their competitors.

That is the principle this manifesto establishes.

The Question Everyone Asks — and Why the Answer Matters​

As analytics platforms mature, a seemingly reasonable question often arises:

If enough data is collected across many businesses, shouldn’t it be possible to benchmark performance?
If hundreds of luxury stores exist across major cities, couldn’t we identify which locations perform best, which layouts convert more, or which footfall patterns signal success?

From a purely technical perspective, the answer appears obvious. With sufficient data, models can be built. Trends can be extracted. Patterns can be learned. Comparisons can be made.

And this is exactly where the problem begins.

What appears to be “industry benchmarking” is, in reality, the reuse of one customer’s operational reality to inform another’s decisions. A luxury store’s footfall dynamics in Paris become training material. A shopping mall’s tenant performance in one city becomes a reference point. An airport’s passenger behaviour becomes a baseline.

Nothing personal is revealed. Nothing identifiable is exposed. Yet something far more valuable is transferred.

The uncomfortable truth is this: benchmarking requires someone else’s data. And when that data represents how a real business performs in a real location, it is no longer neutral. It is proprietary.

This is why the answer is not “we haven’t built it yet”, but “we will not build it.”

Not because it is technically difficult.
Not because it lacks market demand.
But because it violates a fundamental boundary: your data exists to improve your decisions — not someone else’s.

Once this boundary is crossed, the platform no longer serves its customers equally. It quietly turns some customers into data sources, and others into beneficiaries.

That is the line this manifesto exists to draw.

Our Position — Clear and Non-Negotiable​

Data privacy is no longer only about people.
It is about business intelligence, competitive dynamics, and strategic advantage.

While regulations focus on protecting individuals, a far more valuable asset is quietly overlooked:

Customer-owned operational data — and the intelligence derived from it.

This manifesto defines a firm principle:

Your data. Your intelligence. Your AI. For your use only.

Beyond Personal Privacy​

We fully support and enforce the protection of individual privacy:

  • No personal identification
  • No biometric storage
  • No re-identification
  • No surveillance misuse

But this is not enough.

Modern analytics systems observe:

  • Footfall patterns
  • Dwell behaviour
  • Conversion dynamics
  • Queue efficiencies
  • Temporal and spatial correlations

When aggregated over time, this becomes strategic business knowledge — knowledge that can:

  • Reveal demand elasticity
  • Predict revenue performance
  • Expose operational weaknesses
  • Encode location-specific consumer behaviour

This intelligence belongs to the data owner — not the platform, not the vendor, not the ecosystem.

The Core Principle​

Customer-owned data must never be used to benefit other customers.

That includes:

  • Training shared or global AI models
  • Cross-client benchmarking without explicit consent
  • Deriving competitive insights transferable across brands, locations, or operators

A luxury retailer’s data in Paris must never improve a competitor’s performance in Amsterdam.
An airport’s passenger dynamics must never become a template for another airport without permission.

Data leakage does not require data sharing.
Model training itself can be leakage.

What We Will Do​

We will:

  • Build AI models that are logically and operationally isolated
  • Ensure customer data trains only customer-specific intelligence
  • Allow customers to benefit from their own historical depth, not others’
  • Use external, publicly available datasets to enhance context, not replace ownership

External data may include:

  • Weather and climate trends
  • Public economic indicators
  • Population and demographic statistics
  • Transport schedules and infrastructure data
  • Open geographic and mapping datasets

These are contextual enhancers, not substitutes for proprietary data.

What We Will Not Do​

We will not:

  • Sell, trade, or repurpose customer business data
  • Monetise derived intelligence across clients
  • Operate opaque “global learning pools”
  • Treat customer data as a raw material for platform advantage

If a system improves because of your data,
that improvement is yours — not shared, not diluted, not resold.

Why This Matters Now​

AI has shifted the value chain.

The most valuable asset is no longer the algorithm —
it is the longitudinal, real-world operational data that trains it.

Without clear principles:

  • Vendors accumulate asymmetrical power
  • Customers unknowingly subsidise competitors
  • Trust becomes impossible to audit

Compliance alone cannot solve this.
Principles must.

Our Commitment​

This is not a feature.
It is not a configuration option.
It is not a pricing tier.

It is a design philosophy.

If you own the operation, you own the intelligence.
If you own the intelligence, you control its destiny.

That is the standard we choose to uphold —
even when it is less convenient,
even when it limits short-term leverage,
because it is the only model that scales with trust.