The study objective was to assess if a 3-week intervention with the Somnox sleep robot had effects on symptoms of insomnia, somatic arousal, and/or concurrent symptoms of depression and anxiety in adults with insomnia, compared with a waitlist-control group. The participants (n = 44) were randomized to a 3-week intervention with the sleep robot (n = 22), or to a waitlist-control group (n = 22). The primary outcome measure was the Insomnia Severity Index administered at baseline, mid-intervention, post-intervention and at 1-month follow-up. Secondary outcome measures were the Pre-Sleep Arousal Scale, and the Hospital Anxiety and Depression Scale. Additionally, sleep-onset latency, wake time after sleep onset, total sleep time and sleep efficiency were measured the week prior to and the last week of the intervention, both subjectively with the Consensus Sleep Diary and objectively with wrist actigraphy. Mixed-effects models were used to analyse data. The effect of the sleep robot on the participants' insomnia severity was not statistically significant. The differences between the intervention group and the control group on the measures of arousal, anxiety and depression were also not statistically significant, and neither were the sleep diary and actigraphy variables. In conclusion, a 3-week intervention with daily at-home use of the robot was not found to be an effective method to relieve the symptom burden in adults with insomnia.