Automated passive sensing applications and self-reported smart diaries seem to hold promise for the management of anxiety in autism and other mental health conditions. However, passive sensing often struggles with noisy data, ambiguous feedback and weak user agency over the device, whilst self-reporting relies on user-entered data which can be time consuming and cognitively demanding. To address these limitations, we explore a different approach, whereby individuals consciously actuate personal data capture and are in control of it at all times; yet, the interaction solely involves clicking a button, thus avoiding cognitive overload whilst supporting immediate reflection. We call this approach intentive computing. Through our initial investigations we found that conscious interactions cannot only provide real-time relief in anxiety management, but can also function as memory anchors irrespective of the content captured and even prior to data visualization