Personalisation 4.0:
how AI is transforming skin diagnostics

1. Personalisation before AI and after

For decades, "personalisation" in skincare was a marketing promise: products for dry, combination, or oily skin, for under-30s or over-50s, for sensitive or reactive complexions. Useful categories, but fundamentally coarse. The biological reality of skin is far more granular: it varies with the hormonal cycle, the season, stress levels, sleep quality over the past 72 hours, and the humidity of the environment in which one lives.

Artificial intelligence applied to skin diagnostics is making possible — for the first time in a scalable and accessible way — a form of personalisation that reflects this real complexity. No longer predefined categories, but individual biological profiles built on biometric data, updated in real time and translated into routine recommendations that change as conditions change.

The qualitative leap is enormous. Between "dry skin" as a generic label and a skin profile with a specific NP ceramide deficit, excess sebum in the T-zone, irregularities in the circadian skin rhythm, and medium-to-high oxidative stress markers — there is the difference between a one-size-fits-all garment and one made to measure.

"True personalisation doesn't begin in the shop. It begins in the cell."

2. SkinGPT and biometric skin analysis

The term "SkinGPT" entered the beauty tech lexicon in 2025 to describe systems of generative artificial intelligence applied to cosmetic analysis and formulation. Analogously to the large language models that generate text, these systems process biometric skin inputs — high-resolution images, spectrometric data, environmental parameters — to generate precision outputs: diagnoses of the skin's current state, forecasts of its evolution over time, and personalised routine recommendations.

Next-generation biometric analysis technologies go far beyond a photograph of the face. The most advanced systems integrate multispectral data — analysing skin in wavelengths invisible to the human eye — to detect melanin concentration in specific layers, dermal thickness, sebum distribution, vascular patterns, and even early markers of UV damage not yet visible on the surface.

The most innovative aspect is not the diagnostics itself, but its capacity to be adaptive over time: systems like SkinGPT do not produce a static profile, but a dynamic one that updates with every new analysis, tracking the skin's evolution in response to applied protocols and environmental changes. It is the difference between an X-ray and real-time monitoring.

3. Perfect Corp and PHOENIX Pharma: diagnostics in the pharmacy

January 2026 marked a significant turning point in the history of accessible skin diagnostics: Perfect Corp — global leader in AI and AR applied to beauty — announced its partnership with PHOENIX Pharma Italia to bring its Skincare Pro system to the BENU parapharmacy at CityLife in Milan. A collaboration that marks the entry of precision diagnostics into the pharmacy environment, democratising access to consultations previously reserved for specialist clinics.

Through intelligent illuminated mirrors provided by Inovshop, customers access a facial scan that simultaneously analyses eight skin parameters — spots, wrinkles, acne, redness, eye bags, dark circles, radiance, and tone — and returns an individual profile with personalised product recommendations. Not a generic counter consultation, but a biometric output translatable into a specific routine protocol.

PHOENIX Pharma Italia has already declared its intention to expand the solution to further pharmacies, with future developments including CRM integration and evolution towards longitudinal monitoring of each customer's skin profile over time. The pharmacy-as-proximity-clinic model is becoming a reality.

4. Environmental factors and stress: the adaptive routine

The most advanced frontier of AI personalisation concerns not only the diagnostics of the skin's current state, but its correlation with the factors that determine it: air quality, ambient temperature and humidity, chronic stress levels, sleep quality, cyclical hormonal variations. The "adaptive routine" — one that modifies dynamically in response to these inputs — is the next step beyond static diagnostics.

Systems such as the one developed by Perfect Corp are already incorporating environmental data into their models: the UV index of the day, the air quality in the user's geographic area, the season. But the direction of research points towards integration with biometric wearables — heart rate variability as a proxy for stress, sleep quality over recent nights, any hormonal variations — to build recommendations that answer not just "what is your skin like today?" but "why is your skin like this today, and what does it need tomorrow?"

5. Neurocosmetics: when skincare speaks to the brain

Neurocosmetics is the most radically innovative branch of skin personalisation 4.0: the range of cosmetic products formulated to interact with the receptors of the peripheral nervous system, modulate sensory perception, and — through the skin-brain axis — positively influence mood, reduce perceived stress, and improve the user's emotional wellbeing.

The scientific foundation is solid: the skin is not merely a physical organ but a neuroendocrine and immunological one, home to approximately 200 sensory receptors per square centimetre and intimately connected to the central nervous system through a network of nerve endings, neuropeptides, and neurotransmitter mediators. Neurocosmetics act on this network — not "on the brain" directly, but on its cutaneous interface — producing measurable effects on serotonin, beta-endorphins, and cortisol.

"The skin and the brain form from the same embryonic tissue. It is no surprise that they continue to talk to each other throughout life."

6. Functional aromatherapy: the science of olfactory stimuli

Aromatherapy has long suffered from a scientific credibility problem — associated with a new age imagery that obscured its real biological basis. 2026 is changing this perception: research published in Cosmetics has systematised the evidence on bioactive aromatic compounds and their mechanisms of action, paving the way for what is now called functional aromatherapy — clearly distinguishing it from the purely hedonic use of fragrances.

The olfactory pathway is the only one that reaches the limbic system — the brain structure governing emotions, memory, and instinctive responses — without passing through the thalamus, the filter that mediates all other sensory stimuli. An aroma can therefore alter nervous system chemistry before the conscious mind has even identified it. This speed of action makes functional fragrances potentially integrable with AI diagnostics: a system that knows your stress profile could also recommend the olfactory composition best suited to your neurological state at that moment.

Aromatic compounds with the most robust neurophysiological evidence

• 01
  Linalool (lavender)
  The most studied compound in clinical aromatherapy. GABA-A receptor modulation with documented anxiolytic effect in controlled studies. Measurably reduces salivary cortisol levels within 10 minutes of exposure.
• 02
  Limonene (citrus)
  Documented effects on mood improvement and cognitive performance. Japanese studies demonstrated significant improvement in mild depressive states with exposure to citrus fragrances. Also activates cutaneous receptors with a local antioxidant effect.
• 03
  α-Santalol (sandalwood)
  Sedative and anxiolytic effects through GABA receptor modulation. Particularly effective in supporting the transition to deep sleep. Neuroimaging studies show reduced sympathetic nervous system activity with prolonged exposure.
• 04
  β-Caryophyllene (black pepper, cloves)
  Selective agonist of the CB2 receptor in the endocannabinoid system. Anti-inflammatory effects both locally — when absorbed transdermally — and systemically. One of the terpene compounds with the most robust neurophysiological evidence profile.

7. The boundary between personalisation and biological data privacy

Precision personalisation carries a cost that goes beyond the financial: biometric skin data — high-resolution facial images, stress profiles, hormonal variations, ageing patterns — are sensitive biological information. Their management, storage, and commercial use raise privacy questions that the European regulatory framework (GDPR) has not yet fully resolved for this specific category of data.

The fundamental distinction to understand as a consumer is between anonymous in-store analysis — which does not retain the biometric profile beyond the session — and longitudinal profiles with integrated CRM — which build an archive of individual biological data over time. Both have value, but they require very different levels of consent and transparency.

8. How to navigate new technologies without becoming dependent on them

AI personalisation technologies carry an intrinsic risk that echoes the one discussed in relation to wellness more broadly: they can become yet another source of dependence on digital feedback, making the consumer progressively less capable of reading their own skin autonomously and ever more reliant on the algorithm's "diagnosis".

The most mature approach uses AI diagnostics as a starting point and verification tool, not as a permanent oracle. A biometric scan every 2–3 months to track evolution and adjust the routine is an intelligent use of technology. A weekly scan that generates anxiety every time parameters fail to improve is a dysfunctional use of the same technology.

"The best technology is the kind that increases your autonomy — not the kind that creates dependence on its own approval."

9. Building a bespoke routine: from data to practice

A truly personalised skincare routine — one that integrates the principles of AI diagnostics with knowledge of one's own skin and the consistency of JOMO — is built in layers. Not everything requires technology. Not everything can be delegated to the algorithm. Precision personalisation is the meeting point between biometric data and the autonomy of personal judgement.

The 5-layer protocol

• 01
  Reference diagnostics (quarterly)
  An AI biometric scan every 3 months — in a pharmacy, an equipped beauty retailer, or with a dedicated app — to obtain an updated skin profile and verify the effectiveness of the current routine. No more frequently: the skin needs time to show real responses.
• 02
  Constant foundations (daily)
  Regardless of the AI profile, certain elements are invariant: gentle cleansing at acidic pH, basic hydration, morning sun protection. These do not change with the season or biometric data — they are the substrate on which everything else is built.
• 03
  Personalised layer (adaptive)
  The layer that changes with the data: the specific serum targeting issues identified by diagnostics, the antioxidant booster during periods of high environmental exposure, the calming treatment during phases of elevated stress or in the luteal phase of the cycle.
• 04
  Systemic integration (daily)
  The part of the routine that doesn't show in the app but is felt in the skin: omega-3s, probiotics, adaptogens, bioavailable vitamins. The systemic support that amplifies the effectiveness of any topical protocol.
• 05
  Neurocosmetic input (ritual)
  A moment of the day — morning or evening — in which the routine also becomes a sensory ritual: functional fragrances, textures chosen for tactile pleasure as well as efficacy. The component that transforms skincare into self-care in the fullest sense.