What is Human-Centered AI Service Management (HUCAISM)?^®)?

If AI takes over certain services, who is responsible?

Artificial intelligence is transforming how services are created and delivered—faster than traditional service management frameworks and AI governance standards can keep up. It is precisely in this gap that problems arise: AI-powered services that belong to no one, for which no one takes responsibility when something goes wrong.

HUCAISM (Human-Centered AI Service Management) bridges this gap. It is the discipline and operational model that organizations use to design, deliver, and manage AI-powered services in a way that keeps human judgment, human responsibility, and human value at the center. HUCAISM combines the proven practices of service management with the requirements of AI governance into a consistent, operational approach—as a framework from SERVIEW, not as a replacement for ITIL, but as a complement: While ITIL describes how services are organized, HUCAISM describes how they function when AI is involved in decision-making.

The framework comprises 6 principles, 10 domains, and 65 verifiable requirements, supplemented by 3 oversight models, 6 life cycle phases, and 5 maturity levels.

About the HUCAISM Professional Training Course with SERVIEW

Why is HUCAISM needed?

Traditional service management was developed for largely predictable systems. AI-powered services, however, operate on a probabilistic basis: They can produce varying results, change during operation due to changes in models, prompts, or knowledge, and provide answers that are convincingly phrased but factually incorrect.

At the same time, established approaches to AI governance often focus on the AI system itself and its regulatory classification. However, they do not fully address how the specific service is managed, monitored, improved, and, in an emergency, taken offline during day-to-day operations.

HUCAISM brings these perspectives together. It extends service management to incorporate the unique characteristics of AI and translates governance, risk, and compliance requirements into day-to-day operations. This creates a bridge between technical implementation, organizational responsibility, and real-world practice.

Who HUCAISM is for:

HUCAISM is aimed at anyone who is responsible for, designs, operates, or uses AI-powered services and digital products: service managers, executives, consultants, and professionals from IT, compliance, and business units. A technical background is not required. A foundation in service management is helpful—ideally with ITIL, regardless of the version—since HUCAISM is built on the common core of these practices rather than their differences.

Positioning: A Partner, Not a Rival

HUCAISM is entering an already crowded landscape of frameworks, standards, and laws—and explicitly sees itself as a connector, not a replacement:

• ITIL & Service Management: HUCAISM builds on proven practices rather than replacing them. It does not re-teach incident, change, or service level management, but rather describes how these concepts change once AI comes into play. With ITIL 5—which, among other things, will introduce a dedicated module on AI governance in 2026—HUCAISM sees its approach validated: While ITIL 5 defines the “what” and “why” at a framework level, HUCAISM provides the “how” for each individual AI-supported service.
• ISO/IEC 42001: The standard for an AI management system serves as the organizational framework—it establishes the basic principles for how an organization handles AI. HUCAISM is the layer below that, which fills this framework with day-to-day operations.
• EU AI Act: The binding legal framework sets out the requirements; HUCAISM translates these requirements into concrete design and operational decisions. Legal interpretation remains the responsibility of legal experts—HUCAISM describes the operational logic.

If AI takes over certain services, who is responsible?

Artificial intelligence is transforming how services are created and delivered—faster than traditional service management frameworks and AI governance standards can keep up. It is precisely in this gap that problems arise: AI-powered services that belong to no one, for which no one takes responsibility when something goes wrong.

HUCAISM (Human-Centered AI Service Management) bridges this gap. It is the discipline and operational model that organizations use to design, deliver, and manage AI-powered services in a way that keeps human judgment, human responsibility, and human value at the center. HUCAISM combines the proven practices of service management with the requirements of AI governance into a consistent, operational approach—as a framework from SERVIEW, not as a replacement for ITIL, but as a complement: While ITIL describes how services are organized, HUCAISM describes how they function when AI is involved in decision-making.

The framework comprises 6 principles, 10 domains, and 65 verifiable requirements, supplemented by 3 oversight models, 6 life cycle phases, and 5 maturity levels.

The Six Principles: Decision-Making Rules

The principles of HUCAISM are not intended as a list of values, but rather as decision-making rules for moments when two legitimate goals conflict—such as efficiency versus human oversight, or speed versus care. They form the normative foundation upon which the entire model rests:

1. Responsibility cannot be delegated—for every AI-powered service, it must be clear who is accountable. “The AI decided that” is not an acceptable answer.
2. Augmentation before substitution—AI empowers people first, rather than replacing them. Full automation is a conscious choice, not the default path.
3. Transparency in the service, not just in the model—It must be clear how a service arrives at its result: for users, operators, and auditors.
4. Value is created for people—success is measured by the problems solved and the trust built, not solely by throughput.
5. Governance by Design – Regulatory and ethical requirements belong in the design phase, not in the post-design phase.
6. Supervised learning—AI-powered services should be allowed to improve, but only under controlled conditions, with a verified fallback plan in case of an emergency.

The Ten Domains: Where the Model Is Effective

Built upon these principles are the ten domains—the actual areas of focus in which the 65 assessable requirements are located. The focus is on the individual, who serves as the decision-maker, operator, and beneficiary of the service. Four groups of domains are organized around this individual:

Five core domains underpin every AI-powered service:

• KA1 – People & Roles: Who makes decisions, who carries out operations, who is responsible for oversight—and why oversight without the necessary expertise and time is merely a facade.
• KA2 – Practices: How incident, change, service request, and knowledge management change when AI comes into play.
• KA3 – Governance & Compliance: The EU AI Act and ISO/IEC 42001 as Design Parameters Rather Than a Burden.
• KA4 – Value & Experience: Output vs. Outcome – Why Rising Metrics Can Be Misleading.
• KA5 – Technology & Data: Models, Drift, Monitoring, and Recurrence Curves from an Operational Perspective.

Two cross-cutting domains permeate all the others:

• KA6 – Fairness & Bias: Why Every Service Needs a Deliberate, Documented Commitment to Fairness.
• KA7 – Trust & Human Factors: Automation Bias, Trust Calibration, and the Proper Handoff Between Human and Machine.

A specialized field comes into play where AI itself takes action:

• KA8 – Governance of Agent-Based Systems: Limits of Autonomy, Scope of Influence, and the Tested Emergency Shutdown for Active AI.

Two areas of collaboration govern how humans and machines work together to provide the service:

• KA9 – Interaction & Behavior: When AI speaks, remains silent, asks questions, or voluntarily hands over control.
• KA10 – Hybrid Teams & Orchestration: People, AI agents, and automation as a single operational unit, with the Service Manager acting as the orchestrator.

You're familiar with your ITIL practices. But are you aware of the impact of AI?

You’re familiar with ITIL. Incidents, changes, service requests, knowledge management—all familiar territory. But as soon as AI starts playing a role in decision-making for your services, these practices begin to behave differently than you’re used to. Here are a few examples:

• An incident where nothing goes down. Your assistant is available, fast, and runs flawlessly—yet still provides a clearly incorrect answer. Traditional monitoring doesn't pick up on this. Is that already an incident?
• A change that nobody requested. The platform provider updates the system overnight—and suddenly your service behaves differently. Who approves this when nobody was asked?
• A problem that can't be "fixed." Sometimes the problem isn't in a line of code, but in a property of the model itself. What does problem management do when there isn't a bug?
• A configuration item that is not a server. What exactly is being configured when a CI defines the service?
• An SLA that no longer guarantees accuracy. What do you tell a customer when the service can only make an educated guess?

HUCAISM doesn't reinvent your practices—it shows you where you need to rethink your approach once AI comes into play. HUCAISM walks you through each individual practice: which new triggers, categories, and controls are added, and what specific adjustments you need to make.

A model to follow, not a checklist to work through

The principles form the foundation; the domains form the structure built upon it. The model is supplemented by a governance layer (three oversight models) and an operational layer (lifecycle, maturity levels, measurement levels, roles)—all framed by a normative core that definitively establishes what applies. Organizations enter the model at the point where their own pain lies—usually an overlooked incident or a question no one could answer—and from there, they discover what else is involved.

From AI System to Responsible Service: HUCAISM Professional

To properly understand HUCAISM and apply it to specific AI-powered services, a solid understanding of the model is essential. This is exactly where the HUCAISM AI Service Professional training course comes in.

The HUCAISM Professional course teaches the fundamentals of the model, explains its six principles, eight domains, roles, and oversight patterns, and provides practical guidance on how to design, operate, measure, and manage AI-enabled services. It is intended for specialists and executives in service management, AI governance, risk, compliance, and management, as well as anyone responsible for AI-enabled services and digital products.

HUCAISM Professional thus provides a comprehensive introduction to HUCAISM—from structured classification and practical application to preparation for human-centered, transparent, and auditable AI operations.