The DCI AI Brain is a knowledge-retrieval system.
Not a chatbot.

The DCI Brain does not generate responses from a general language model. It retrieves, matches, and reasons across the DCI formula knowledge graph — the same rigorously classified, professionally reviewed entries in the Formula Library. Every recommendation is drawn from a documented formula intelligence entry, not generated from statistical patterns.

It makes recommendations. You make decisions. The six non-negotiable human review rules below are not guidelines. They are non-negotiable. They appear on every result.

Open Formula Assistant

What the DCI Brain IS

  • A formula recommendation system — retrieval-augmented reasoning over the DCI formula database
  • A hair diagnosis guide — structured cuticle, elasticity, and structural risk assessment
  • A consultation coach — question sets and archetype flags derived from Vol 6
  • A confidence scorer — HIGH, MEDIUM, or DEVELOPING on every output

What the DCI Brain IS NOT

  • A replacement for physical assessment at the chair
  • A guaranteed clinical outcome — every recommendation requires professional verification
  • A general AI assistant — it does not answer questions outside its knowledge domain
  • A product recommendation engine — it recommends formulas, not brands

Four DCI AI tools.

Each tool is a structured interface to the DCI knowledge graph. Each one returns structured output — not a conversation.

Live

Formula Assistant

Six-variable structured input → formula recommendation with confidence score, rationale, alternatives, and maintenance plan.

Free preview available
Open
Live

Hair Diagnosis

Guided cuticle state, elasticity, and structural risk assessment. Returns protocol recommendations.

Free preview available
Open
Live

Consultation Coach

AI-generated consultation question set by client type. Surfaces expectation management flags.

Professional plan
Sign in
Beta

Case Builder

Enter a service narrative; AI populates the DCI case format fields for review and submission.

Coming soon — beta
Planned

Visual Hair Diagnosis

Camera-based level and tone estimation. Upload or capture a photo; AI returns level estimate and canvas condition flags.

On the roadmap

Free preview: 5 Formula Assistant queries per month without a subscription. Unlimited access, all tools, on the Professional plan.

Six non-negotiable human review rules.

These rules are displayed on every AI result, in every interface. They cannot be dismissed, minimised, or hidden. They are the permanent boundary between AI assistance and clinical responsibility.

Six Non-Negotiable Human Review Rules

The AI returns recommendations. The professional makes decisions. These six situations require independent clinical assessment regardless of confidence score.

1. Structural integrity must be assessed in-person

Absolute

If elasticity result = none (state 3) or if the client reports significant breakage: the AI recommendation does not apply. Physical assessment supersedes AI output in all structural integrity cases.

2. Allergy and sensitivity history requires direct questioning

Absolute

The AI has no access to the client's medical history. PPD sensitivity, persulfate sensitivity, and prior reactions require direct in-consultation questioning before any chemical service.

3. Canvas level must be physically confirmed before toning

Absolute

The AI accepts a canvas level as input. It cannot verify that the stated canvas was achieved. The professional must physically confirm the canvas under salon lighting before applying any toner formula.

4. Prior chemical service history must be complete

AI recommendations assume the chemical history input is accurate and complete. Undisclosed services (relaxers, prior bleach, metallic dye) can produce dangerous interactions. The professional is responsible for establishing a complete history.

5. Developer selection requires independent professional verification

The AI recommends a developer concentration based on the stated canvas and target. The professional must independently verify this selection is appropriate for the specific client's structural condition on the day of service.

6. Correction services require full visual assessment

AI correction recommendations are route guides, not protocols. Correction services involve variables (prior deposits, uneven porosity, competing undertones) that cannot be fully captured in a six-variable input. Professional visual assessment is required.

Absolute — AI output is blocked. Do not proceed.
Required — independent professional assessment mandatory.

Professional Assessment Required

This is a recommendation. Professional assessment and judgment are required before application. Canvas level, structural condition, and chemical history must be assessed physically at the chair.

What confidence levels mean clinically.

Every DCI AI recommendation carries a confidence level. Understanding what each level means is part of the professional responsibility of using the tool.

What Confidence Levels Mean

CONFIDENCEHIGH

Strong formula match

Complete input provided. The formula is drawn from a well-documented case category. Canvas requirement confirmed. Developer selection is straightforward. Professional review is still required — but the recommendation is grounded in strong evidence.

CONFIDENCEMEDIUM

Partial input or developing evidence

Either the input is incomplete (some variables estimated) or the formula evidence base is professional consensus rather than documented case evidence. The recommendation is directionally sound — but requires closer professional assessment before application.

CONFIDENCEDEVELOPING

Minimal input or inference-based

The input is minimal, or this formula territory relies heavily on professional inference rather than documented evidence [S4]. The recommendation should be treated as a starting point for the professional's own assessment — not as a clinical guide.