Almost half of GP practices are using AI-supported digital healthcare applications; a year earlier, this was only about one in five practices. However, virtually the entire growth is attributable to a single application that the patient never sees. Resistance to chatbots in patient contact remains undiminished.

 

This is evident from a repeat study by Nivel, which was distributed on October 6, 2025, to over 4.500 general practitioner practices in the Netherlands. Over 800 practices completed the questionnaire in full. Nivel notes, however, that the results may not be representative of all general practitioner practices. The thinking is that practices already working with AI are more likely to complete such a list.

The survey comprised two blocks. First, questions were asked about five general digital applications: online appointment scheduling, digital triage, video calling, telemonitoring, and digital mental healthcare. Additionally, for the second year in a row, questions were asked about six AI-supported applications: triage, chatbot or digital doctor's assistant, voice-activated reporting, diagnostics, preventive care, and personalized treatment plans.

Digital mental healthcare is used by 85% of practices, and online appointment scheduling by 78%. Telemonitoring (57%), digital triage (45%), and video calling (40%) lag behind. However, the frequency of use presents a different picture: online appointments are used daily by 64% of practices, whereas the daily use of digital mental healthcare, telemonitoring, and video calling stands at 18%, 10%, and 3%, respectively. Digital triage is used daily by 22%, but 55% of practices do not make use of it. This application was surveyed for the first time in 2025, so a comparison with 2024 is missing. For the remaining four applications, the picture is virtually identical to last year.

Growth in AI usage, but little deployment of chatbots

46% of practices deployed at least one of the six AI applications in 2025. Broken down: voice-controlled reporting 36%, triage 16%, diagnostics 7%, chatbot or digital doctor's assistant 6%, preventive care 4%, and personalized treatment plans 2%. Incidentally, some of the applications may be rule-based AI.

In addition to practices already using AI, a third to half indicate that they do not use an application but would like to. Half (50%) say this about AI support in preventive care, followed by triage (45%), voice-activated reporting (41%), and personalized treatment plans (41%).

However, the opposite occurs just as often. Slightly less than half of the practices have explicitly stated they have no need for an AI chatbot or digital doctor's assistant to answer patient questions (47%) or for AI support with diagnostics (42%). In the case of the chatbot, opposite that 47% is a group of 35% who would like the application, while 12% are undecided.

Regarding personalized treatment plans, 21% of practices do not know whether the application is being used or if there is a need for it. For triage and voice-activated reporting, these figures are 9% and 8%, respectively. The further removed the application is from daily work, the more uncertainty hangs over the organization.

'Impersonal'

General practitioner practices cite various reasons why they do not feel the need for one or more AI-supported digital applications. This primarily concerns the use of a chatbot/digital doctor's assistant to answer patient questions, or AI support for diagnostics. Reasons frequently cited include the impersonal nature of using AI or the fact that it does not (yet) offer added value to the practice. Several practices also indicate that the AI-supported digital care application(s) are not necessary in their practice, are insufficiently reliable, do not yield time savings, or increase workload. Comments also mention that chatbots are not suitable for the (aging) patient group or do not align with the desire to provide personal care.

Little knowledge

Practices that do have a need to use one or more AI-supported digital healthcare applications, but do not use them, state that they lack sufficient time or financial resources. They also often have priorities other than the implementation of AI-supported applications. Other mentioned bottlenecks are that, in their view, the (desired) AI-supported applications are still insufficiently developed, are not available, or cannot be integrated into their General Practitioner Information System (HIS).

In addition, insufficient knowledge of and unfamiliarity with the capabilities of AI-supported applications are often cited as reasons why people do not use them (yet).

All in all, an upward trend in the use of AI in general practice seems to be emerging. The authors of the Nivel report also refer to the annual study on the use of AI applications in the hospital sector by the AI ​​Monitor Ziekenhuizen of M&I/Partners. This shows that the deployment of AI has increased, but mainly on a small scale. In hospitals, too, AI is used relatively little for communication (including chatbots), certainly when compared to other applications such as in diagnostics or healthcare administration and billing.

The full study can be downloaded here.

(Ziptone/Erik Bouwer)

Also read:

Pushing boundaries in customer contact in healthcare

Featured, Research, Technology