Four months with Isaac in everyday clinical practice
Updated: 23 hours ago
What Kinderarzthaus learned from using Isaac
A preventive check-up, a longer conversation with parents or a psychosocially demanding consultation all require attention. At the same time, the essentials must be documented completely and traceably.
Kinderarzthaus therefore wanted to find out whether AI-assisted documentation actually reduces the workload in everyday practice and whether doctors would voluntarily keep using such a solution after several months.
During the evaluation period, 22 doctors used Isaac to create around 5,300 clinical notes. In addition, 21 doctors took part in a structured survey.
The results paint a nuanced picture. Participants most consistently reported having more attention for the child and parents. Time savings arise mainly in longer and more complex consultations. At the same time, the pilot showed where documentation quality, personalization and technical workflows still need to improve.
The pilot at a glance
22 doctors in the pilot | ~5,300 clinical notes |
18 of 22 active in at least three months | 21 survey participants |
For many, the pilot turned into a regularly used working tool
The pilot started in April 2026 with a smaller group. From June onward, between 20 and 22 doctors actively used Isaac, depending on the month.
Use remained voluntary throughout the pilot. No one was required to use Isaac for a certain number of consultations or for specific types of appointments.
18 of the 22 participants were active in at least three different months. For the vast majority, use was therefore not limited to the introduction or a few test days.
Different usage patterns emerged. Some doctors used Isaac for a large share of their consultations. Others used the assistant specifically for preventive check-ups, longer medical histories, and psychological or psychosocial conversations.
For very short appointments, on the other hand, manual documentation could be just as fast or faster.
When assessing an AI assistant, the number of uses alone is therefore not decisive. What matters is whether doctors voluntarily use the system over a longer period and in which situations they see added value in it.
More attention for the child and parents
19 of the 21 respondents said that Isaac led to better patient contact, more eye contact, better opportunities for observation or more available attention during the consultation.
Instead of constantly typing notes into the computer during the conversation, they could focus more on the child, the parents and the dynamics of the conversation. This benefit was mentioned particularly often for longer, psychological or psychosocially demanding conversations.
This presence is especially relevant in pediatrics. Observing the child, non-verbal signals and interaction within the family are often part of the consultation. Less typing during consultations creates more room for this.
One doctor pointed out, however, that writing her own notes helps her structure her thoughts and make sure that important points are not forgotten. This is also part of the overall picture: ways of working differ, and not everyone benefits in the same way.
More eye contact. Less typing in parallel. More presence during the consultation.
Time savings depend heavily on the consultation
Experiences with time savings were less consistent.
Some doctors estimated their time savings at 25 to 30 percent. Several mentioned five to eight minutes for suitable consultations. One doctor spoke of at least one hour of relief over a working day.
Others noticed no clear time savings. Reasons included longer conversations, the need to shorten or correct the documentation, and technical delays. For short consultations, manual documentation could be faster.
The impact on evening documentation also varied. Eleven of 21 respondents reported less after-hours documentation or work at home, five saw no significant change and two tended to see additional effort due to checking and correcting. For the rest, the effect was mixed or the question was not answered directly.
A blanket promise such as “Isaac saves five minutes per consultation” cannot be derived from this.
Rather, the experiences show that time savings depend on the type of consultation, personal ways of working and the amount of editing required.
The most active users want to keep using Isaac
To assess the pilot, it is worth looking at the survey and actual usage together.
The six most active users account for more than half of all uses within the surveyed group. All six want to keep using Isaac.
The positive assessment therefore also comes from people who used Isaac intensively in everyday practice over many weeks.
Overall, 20 of the 21 respondents want to continue using Isaac voluntarily. Some of them want to use the assistant deliberately and selectively, especially for longer or more complex consultations. One person was mostly critical and did not clearly agree to continued use.
At the same time, intensive users in particular made concrete suggestions for improvement. This is important for interpreting the results: regular use does not mean that weaknesses are overlooked. Quite the opposite. Anyone who uses a system often recognizes very precisely where it makes everyday work easier and where there is still friction.
Where Isaac needs to become more reliable
The pilot also showed areas in which Isaac and the related workflows still needed to improve. One challenge lay in Kinderarzthaus’s existing workflow. Depending on the organ examined, the doctors activate specific text templates by voice command. This process first had to be practiced and aligned with working with Isaac.
There was also room for improvement in the form and style of the documentation. Several respondents felt that the texts were sometimes too detailed and wanted more precise medical keywords. They also wanted more personalization: corrected wording and individual documentation preferences were not retained permanently.
This feedback went directly into further development. With the newly introduced memory function, Isaac is intended to take corrections and personal preferences into account more consistently in the future. How well the function performs in everyday practice and whether it reduces the editing effort will become clear with further use. At the same time, the experience with the text templates shows how important it is to closely align the solution with the respective clinical workflow.
Good documentation support must fit the existing workflow
The pilot also showed how strongly the benefit of an AI solution depends on existing workflows.
At Kinderarzthaus, specific text templates are used depending on the organ system examined. This established process had to be aligned with the AI-generated documentation.
That sounds like a detail, but it is crucial in practice. A documentation solution does not work in isolation. It has to fit existing templates, habits, practice software and personal documentation styles.
Here, too, there was clear feedback from the pilot. Several doctors wanted shorter texts, less repetition and closer adaptation to their own way of working. Some prefer continuous prose, others concise bullet points. The desired level of detail also depends on the consultation.
Saipient introduced a memory function during the pilot to take individual corrections and documentation preferences into account more consistently. How well this reduces the editing effort will become clear with further use.
There was also technical feedback on recordings, transfers to the practice management system, templates and processing times. Several of these points already improved during the pilot. For long-term use, a stable and traceable workflow remains a prerequisite.
Expectations go beyond documentation
During the pilot, requests also emerged for further administrative support, for example with referrals and doctor’s letters, summarizing extensive patient records, or billing and forms.
These functions were not part of the evaluation to date. However, the feedback shows that the need goes beyond documenting conversations.
The same principle applies to such extensions: Isaac can structure information and create suggestions. Review, approval and medical responsibility remain with the professionals.
What is clearer after four months
The pilot does not deliver a simple promise along the lines of “X minutes saved per consultation”. Consultations, ways of working and personal preferences differ too much for that.
Nevertheless, three findings clearly emerge:
More presence: Many doctors report that they document less in parallel during the consultation and can focus more on the child and parents.
Selective time savings: Relief arises mainly where conversations are longer, more complex or more documentation-intensive.
Adoption requires suitable workflows: Doctors use Isaac in different ways. Documentation style, templates and existing processes must therefore be taken into account.
Voluntary use over several months and the strong intention to continue show that Isaac has become a working tool that many participants actually use.
Kinderarzthaus will continue to use Isaac. Use remains voluntary and is guided by the consultations in which it creates real added value for the individual doctor.
“Our most important insight from the pilot project is that the benefit of Isaac is individual and depends heavily on the reason for the consultation. Especially in long and complex conversations, Isaac often creates not only time savings but, above all, more presence for the child and parents. Instead of documenting in parallel, we can keep our eyes on the person in front of us. Isaac was already being further developed during the pilot phase based on our feedback. We have therefore decided to make Isaac available to all our doctors for voluntary use and to continue its development together with Saipient.”
Dr. med. Tim Brennemann, Chief Medical Officer Kinderarzthaus AG
Try Isaac in your own practice
We are happy to share what we learned from the pilot with Kinderarzthaus and to discuss how you can trial Isaac in your practice in a targeted way and evaluate it based on concrete use and feedback.
