I would argue that a digital intervention targeting behavior change should be evaluated no differently than human-interaction based studies. That said, a couple of additional markers may be studied concerning ease of use, perceived value in the interaction, and logistics.
I am attaching links to a couple papers I wrote a while ago about behavior change methods for disease management. These approaches should be implemented in the intervention, whether by human or device.
I hope this helps
Ariel
Article Disease Management Interventions II: What Else Is in the Black Box?
Article Disease Management Interventions: What's in the Black Box?
These reference articles discussed about quality criteria of digital health information, some are not really specific but maybe can help to solve your problem.