Unless you have an SLA for your services response times, the Dynatrace AI will create a problem with an increase in response time from the baseline. If you have SLA's you will need to change the thresholds to fixed thresholds. You can see the Anomaly detection settings for each service (Service -> Settings -> Anomaly Detection) or globally (Settings -> Anomaly Detection -> Settings).
1. You need to tag the services you want to be alerted on - See more on tags here
2. Create an Alerting Profile Based on this tag - See more on Alerting Profiles here
3. Create the problem notification and set the Alerting Profile to the one created as explained above - See more on this here
Hope This Helps