Involve the right expertise

All phases of procuring, implementing, and using AI systems in healthcare should be approached with interdisciplinary collaboration.

To clarify responsibilities, involvement and planning, a responsibility matrix, often referred to as a RACI matrix, can be a useful tool for procurement teams [20][21]. Such a matrix is used to describe different levels of involvement and responsibility that different roles should or can have, depending on the task.

Table 1
The table is an example of a RACI matrix for a procurement process. It shows which roles can be part of a procurement team and what responsibilities they have. R: Responsible (performs the task, can be shared between several roles/people), A: Accountable (has the final responsibility, there can only be one per phase), C: Consulted (someone who is consulted before the final decision is made, requires two-way communication) and I: Informed (someone who is only informed of the decision, one-way communication)

Role

Phase 1

Describe the challenge and assess whether AI is the solution

Phase 2

Map AI systems on the market

Phase 3

Assess your organization's ability to adopt an AI system

Phase 4

Procure an AI system

Phase 5

Implement and quality assure an AI system

Phase 6

Operate and manage an AI system

Project management

I

RA

A

R

A

 

User and/or domain expertise (e.g. clinician)

RA

R

R

C

C

C

IT/enterprise architect

I

R

R

 

 

 

Subject-specific expertise (e.g. radiation protection)

C

C

R

R

R

C

Ethical competence

C

 

C

 

 

C

Citizen/patient/service recipient

I

 

I

 

I

 

Legal expertise

C

C

C

C

C

C

Privacy, information security

 

 C

C

R

R

R

Data management

 

 

R

C

C

C

Analytical expertise

 

R

I

I

R

C

Procurement expertise

 

C

 

R

 

 

Solution architect

 

 

C

R

C

 

Test lead

 

I

I

C

I

 

Operations and management

 

 

C

 

C

R

AI superuser (for example, an "AI radiologist" or "AI medical physicist")

 

 

 

 

C

R

System manager

(IT supplier)

C

I

C

C

I

A

System owner (business management)

C

I

C

A

I

I

For a more detailed description of the roles, see AI fact sheet 0 [22].

 

 

[22] AI fact sheets will be published on Kunstig intelligens 

Last update: 23. mai 2025